Rehabilitation Management of Neck Pain—Development of a Diagnostic Framework Based on the Pain and Disability Drivers Management Model

ABSTRACT

Rationale

Neck pain is a major cause of disability worldwide, and current rehabilitation strategies show limited effectiveness. Subgrouping patients by their primary pain and disability drivers can help tailor treatments. At this end, the Pain and Disability Drivers Management (PDDM) was developed and has demonstrated preliminary effectiveness in the management of low back pain. Nevertheless, the PDDM model was only validated for this population. Adapting this framework to patients with neck pain would provide a more global view of the patient's experience of pain and support a genuine biopsychosocial intervention.

Aims and Objectives

The aim of this study was to develop and validate the content of the PDDM model for patients living with neck pain.

Methods

Through a modified DELPHI study design, participants with clinical and research expertize in rehabilitation of neck pain were invited to participate. A questionnaire was developed using literature reviews and endorsed by a steering committee. The relevance of each element of the newly adapted model was evaluated on a 4‐point Likert scale. An item reached consensus if it obtained the predefined threshold of > 78% “relevant” and “very relevant.” Participants left comments on terminology and recommended items to add in early rounds. Quantitative and qualitative analyses were performed.

Results

An invitation was sent to 1650 potential participants, from which 155 accessed the survey, 64 completed the first round and 55 the second round. A total of 70 elements met consensus and were distributed across six domains: “Nociceptive pain drivers”, “nociplastic pain drivers,” “drivers associated with neuropathic pain”, “comorbidity drivers”, “cognitive‐emotional drivers” and “environmental or lifestyle drivers, and social determinants of health.”

Conclusion

Through a modified DELPHI study, the PDDM model was updated and adapted to people with neck pain. Subsequent steps include clinical integration and measures of efficacy when used for assessment/treatment.

1. Introduction

Musculoskeletal disorders are a major cause of disability, worldwide [1]. Among these, neck pain (NP) is a major contributor to years lived with disability [2]. It is estimated that the number of patients living with NP will increase by 32.5% by 2050 [3], which will most certainly elevate the already high socioeconomic costs of this disease [4]. Effective rehabilitation management strategies are thus needed to address these challenges.

Current rehabilitation guidelines for NP include reassurance, advice to remain active, education, exercises, and manual therapy [5]. However, these interventions demonstrated heterogeneous effect sizes for improving pain and disability [6, 7, 8, 9, 10]. These mixed results may be explained by the fact that the profile of people living with NP is heterogeneous; each person has unique characteristics that can moderate outcomes. A promising way to improve treatment efficacy is to identify subgroups of patients suffering from the same health condition and provide them personalized treatment—namely stratified care [11].

Classification systems were developed to propose specific treatments for subgroups of patients to personalize care options [12, 13, 14]. Despite their apparent advantages, classification systems lacked efficacy in improving outcomes in the rehabilitation care of musculoskeletal disorders, including NP [15]. One potential explanation is that these classification systems tend to prioritize biological factors over psychosocial factors [14], which may downplay the importance of psychological and social influences to NP [16].

To address the limitations of classification systems, Tousignant‐Laflamme et al. proposed a diagnostic framework—namely the Pain and Disability Drivers Management (PDDM) model [17]. This PDDM model proposes to assess the patient's drivers of pain and disability under five domains: (1) Nociceptive pain drivers, (2) nervous system dysfunction drivers, (3) comorbidity drivers, (4) cognitive‐emotional drivers, and (5) contextual drivers [17]. This model was originally proposed in 2017 [17] and its content was validated by experts [18]. Following this validation, the authors developed tools to facilitate its clinical integration [19, 20] and tested its effectiveness with patients suffering from low back pain [21, 22, 23]. This type of framework offers the opportunity to have a more comprehensive view of a patient's pain experience and supports the integration of a meaningful biopsychosocial intervention.

The PDDM has been developed and validated for the specificities of people with low back pain [14]; consequently, there was a need to adapt this model to the characteristics of patients with NP, as conditions associated with NP have unique features that differentiate it from low back pain (e.g., headache, cervical myelopathy, whiplash) as well as specific prognostic factors [24, 25]. The aim of this study was to develop and validate the content of the PDDM Model for people living with neck pain (PDDM‐NP).

2. Methods

2.1. Study Design

We conducted a modified DELPHI study and reported our findings using Guidance on Conducting and Reporting Delphi Studies was used to report the results (CREDES) [26]. This technique is a consensus method using a series of questionnaires to collect data from a panel of selected participants [27]. It is labeled “modified” since we used a pre‐defined questionnaire during the first round, instead of open‐ended questions [28]. Another distinction is that participants were not given the opportunity to reconsider their decision, which differentiates this approach from the classic Delphi method [29]. This technique was previously used to validate the content of the original PDDM [18]. This study design ensured the relevance and completeness of the model's proposed items, considering the need to determine content validity through the consensus of experts. Ethics approval was granted by the Ethics Review Board of the Clinical Research Center of the CHUS (project # 2024‐4933).

2.2. Selection of Experts

Participants were selected through maximum variation purposive sampling [30], which is consistent with the Delphi methodology. Study participants are selected based on their level of competency within the specialized area of knowledge [28]. Participants needed to be considered “experts” in the pain and disability management of musculoskeletal disorders. To be considered, potential participants must satisfy one of the following five criteria: (1) Obtain a graduate certificate in pain management from the Université de Sherbrooke, Québec, Canada; (2) identify as an expert in the field of neck pain according to the https://expertscape.com/ platform (all experts listed on the first page of the “neck pain” domain, n = 61); (3) hold a diploma (or diploma candidate) in Mechanical Diagnosis and Therapy; (4) be a fellow of the Canadian Academy of Manipulative Physiotherapy; or (5) a rehabilitation clinician who has completed training on the PDDM Model for low back pain [17]. Participants had to speak either French or English, be able to take part in three rounds of questionnaires within 2 months and have Internet access to complete the questionnaire on a smartphone or computer. As compensation for their time to complete the survey, participants could choose to receive a $10 gift card once all rounds of the survey had been completed.

2.3. Recruitment

Only potential participants who met the inclusion criteria described above were invited to participate in the study by email. Potential participants from Mechanical Diagnosis and Therapy, the Canadian Academy of Manipulative Physiotherapy and graduates in pain management from the Université de Sherbrooke were solicited by email by these organizations. Potential participants from other groups (PDDM and experts) were contacted directly by the research team.

2.4. Sample Size

A consensus sample size for a DELPHI study is lacking [31]. Based on our experience in validating the PDDM for patients with low back pain, we estimated that we had to send > 1000 invitation emails to “experts” and anticipate a 5% response rate, or ±50 participants. This number is more than sufficient to obtain a variety of perspectives [31].

2.5. Development of the Preliminary PDDM‐NP

To develop a preliminary version of the model, a steering committee consisting of a team and external researchers (co‐authors) was generated. The development of the PDDM‐NP followed an iterative process. We performed a systematic review to identify and evaluate the quality of prognostic factors specific to neck pain and associated conditions (i.e., trauma‐related, cervical radiculopathy) [32]. We extracted characteristics useful to discriminate between the different pain mechanism categories from the recent publications performed by Shraim et al. to inform the domains considered as drivers of pain [33, 34]. We broadly addressed social determinants of health, as growing evidence suggests they may have a significant impact on outcomes [35]. At this end, we used the framework suggested by the US Department of Health and Human Services [36] and the work from the World Health Organization [37] (See Supporting Information 1).

With these considerations, two co‐authors (TG, YTL) built a preliminary version of the PDDM‐NP which was then presented and vetted by the steering committee. This adapted version retains the principle of categories, with category A items that can be addressed by physiotherapists and category B items that are partially addressable and require multidisciplinary management or referral. The steering committee proposed adjustments to its content, terminology, and justifications through an iterative process involving meetings and email exchanges. This process resulted in the development of a preliminary version of the PDDM ‐NP. This version was comprised of 62 elements integrated into six domains: “Nociceptive pain drivers,” “nociplastic pain drivers,” “drivers associated with neuropathic pain,” “comorbidity drivers,” “cognitive‐emotional drivers” and “environmental or lifestyle drivers, and social determinants of health.”

2.6. Delphi Method Procedures

The online platform Limesurvey (https://www.limesurvey.org/) was used to host the questionnaire. The survey was available in French and in English. Three rounds were planned for this DELPHI study. Participants had 3 weeks to complete each round of the survey. A weekly email reminder was sent to participants to increase the response rate.

2.6.1. Round 1

The first round began with a short video introducing the PDDM‐NP and the elements to be validated (domain by domain) (available on https://go.screenpal.com/watch/cZnIoJVdgST). Participants were asked to answer a few questions to ensure that they had adequately integrated the content of the video. An online document was made available to participants, containing the rationale for each domain, along with examples of operationalization. Afterwards, participants were asked to evaluate the relevance of each of the model's proposed items via a 4‐point Likert scale (“highly relevant,” “relevant,” “not relevant,” or “highly irrelevant”) where the middle point was absent to avoid neutral answers. A field was left open to allow participants to propose new elements or terminology suggestions for each domain of the model (during round 1 only). For each item, the percentage of “relevant” and “very relevant” responses was calculated. To reach a consensus, it was predetermined that an item had to obtain at least 78% or more of a “relevant” and “highly relevant” scoring. The 78% threshold was retained, since it corresponds to the item‐level content validity threshold, where the risk of agreement by chance is minimal, regardless of sample size [38]. Items with a “relevant” and “highly relevant” scoring between 25% and 77% were not considered consensus and were re‐evaluated in a second round. Finally, items with less than 25% “relevant” and “highly relevant” were immediately discarded. These cut‐offs were defined from the DELPHI survey previously conducted on the model [18, 39].

2.6.2. Round 2

For the second round, a new questionnaire was elaborated, including items that had not reached consensus during the first round, the new items proposed by the participants, and the terminology modifications for certain elements suggested by the participants. For this round, items and modifications with a “relevant” and “highly relevant” scoring 78% or more were retained in the final model. Items with less than 78% of consensus were discarded. Items proposed in the first round and scoring between 25% and 77% were re‐evaluated in the third round.

2.6.3. Round 3

In the case where elements proposed by our participants in the first round did not achieve consensus during the second round, they were then re‐evaluated during the third round. The third round was therefore optional. For this round, items with a “relevant” and “highly relevant” scoring of 78% or more were retained in the final model. Items with less than 78% of consensus were discarded.

2.7. Analysis

Analysis was limited to fully completed surveys since we wanted to receive feedback on the model to ensure consistency, and because it is important that participants maintain their involvement until the end of the process [28]. Quantitative analyses after each round were performed, as detailed above. Statistical analysis was performed with R (version 4.3.2). For qualitative analysis, the comments left by participants during the first round were analyzed by two authors (TG, YTL). Each of the comments was classified as “new item to be evaluated in next round,” “general comment” or “terminology modification.” This process was then validated by the steering committee.

3. Results

3.1. Participants

An invitation email was sent to 1650 potential participants. The survey was accessed by 155 participants (invitation view rate: 9.4%), 64 completed the first round (41% completion rate), and 55 completed the second round (14% attrition from round 1) [40]. Of the 91/155 who did not complete the questionnaire, 44/91 (48%) did not consent to participate, 21/91 (23%) did not respond to the mandatory question regarding follow‐up to the video presentation, and 26/91 (29%) did not complete the entire questionnaire (Figure 1). Participants' characteristics are shown in Table 1.

Figure 1.

Participants’ flow in each round.

Table 1.

Participants' characteristics.

Variable		Participants
Numbers of participants		n = 64	Percentage
Age	< 30 years	3	5%
	30–39 years	19	30%
	40–49 years	22	34%
	50–59 years	12	19%
	≥ 60 years	8	13%
Sex	Female	30	47%
	Male	34	53%
Gender	Woman	30	47%
	Man	34	53%
Country	Australia	4	6%
	Belgique	1	2%
	Canada	30	47%
	Czech Republic	2	3%
	Danmark	2	3%
	Deutschland	1	2%
	France	2	3%
	Greece	1	2%
	India	3	5%
	New‐Zealand	1	2%
	Norway	1	2%
	Spain	1	2%
	United Kingdom	1	2%
	United States of America	14	22%
Profession	Researcher	2	3%
	Physiotherapist	55	86%
	Chiropractor	1	2%
	Nurse	1	2%
	Physical rehabilitation technician	2	3%
	Other	3	5%
Experience (years)	< 5 years	2	3%
	5–9 years	9	14%
	10–14 years	13	20%
	14–19 years	12	19%
	> 20 years	28	44%
Background	Experts from expertscape. com	8	13%
	Pain and Disability Drivers Management group	9	14%
	Mechanical Diagnosis and Therapy	32	50%
	Pain management course	6	9%
	Canadian Academy of Manipulative Physiotherapy	9	14%
First language	English	49	77%
	French	15	23%

3.2. Results of the First Round

Of the 62 initial elements of the PDDM‐NP, 60 obtained a score above the predefined threshold of 78% and were integrated into the final model (interval of agreement: 80%–100%). A consensus was reached on all the rationale for the A and B categories (interval of agreement: 97%–100%) (see Supporting Information 2).

This first round validated 9 of the 10 elements for the nociceptive pain driver's domain, and no new element was suggested. The only item that did not reach consensus was the Inflammation marker 3 — Pain is described using terms such as: Swollen, stiffness, crackling, dull aching or throbbing” which obtained 72% of agreement). For the nociplastic pain drivers, the seven proposed elements reached consensus, and two new elements were suggested (“non‐consistent mechanical patterns of pain reproduction (e.g. inconsistent reproduction of pain by movement” and “Widespread pain”). For the drivers associated with neuropathic pain, the eight proposed elements reached consensus, and two new elements were suggested (“Neurological symptom modification (e.g., decrease of paresthesia during cervical movement or clinical testing)” and “Gait ataxia”). For the comorbidity drivers, 10 of 11 elements reached consensus, except one (“overweight” obtained 66% of agreement), and three new elements were proposed (“Chronic primary pain (e.g., fibromyalgia, chronic fatigue syndrome, irritable bowel syndrome)”; “Symptoms associated with vestibular disorders” and “Endocrine disorders (e.g., diabetes, hypothyroidism)”). For the cognitive‐emotional drivers, all 14 elements reached consensus, and four new ones were proposed (“readiness to change”, “perceived injustice”, “feelings of guilt” and “health‐seeking behaviors”). Finally, for drivers related to environment or lifestyle drivers, and social determinants of health, the 12 elements reached a consensus, and no new element was proposed. The complete results of the first round are available in Supporting Information 3.

3.3. Results of the Second Round

The second‐round survey was comprised of 13 items, namely the two elements that did not reach consensus during the first round and 11 elements, newly proposed elements by the participants. Minor terminology modifications proposed by participants in the first round were also assessed during this round (e.g., remove the example “stroke” from the item “Presence of central neurological pathology (e.g., stroke, cervical dystonia)”). After this second round, 11/13 items reached consensus and two were not validated (Inflammation marker 3 — Pain is described using terms such as: Swollen, stiffness, crackling, dull aching or throbbing and overweight respectively obtained 74% and 58% of agreement). All the minor terminology modifications were validated (reached consensus) by our panel of participants (see Supporting Information 4). The item “gait ataxia,” already included as evidence of myelopathy, was merged with the latter to avoid redundancy. As all elements proposed during the first round met with consensus during this second round, there was no need for a third round. The final version of PDDM‐NP included 70 elements that reached consensus within the six domains (Tables 2, 3, 4 and Figure 2).

Table 2.

Drivers of pain according to the Pain and Disability Management Model for people living with neck pain (PDDM‐NP).

Domain 1 – Nociceptive pain drivers
Category A: Mechanical pain pattern, reproducible and responsive to symptom modification procedures (e.g., centralization of symptoms following repeated movements, pain reduced through postural modification).		Category B: Mechanical pain pattern, reproducible, but unresponsive to symptom modification procedures OR mechanical pain of systemic inflammatory origin.
Patient presents centralization of pain during examination	95%	No biomechanical modification can reduce the patient's pain	89%
During the assessment, certain biomechanical modifications allow movements to be performed with less pain (e.g.,: Changes in posture, shoulder elevation, NAGs or SNAGs, etc.)	100%	Pain is reproduced from the initiation of a movement (during the movement and often worse end‐range) and in several planes of movement (without a clear and constant pattern of pain)	87%
The patient presents biomechanical characteristics that predict they will respond to a specific approach (e.g., symptomatic response indicating a positive response to directional preference exercises).	98%	Inflammation marker — Onset: Typically, within 2 weeks or recent flare‐up of a chronic condition	89%
		Inflammation marker — Pain usually resolves within a timeframe that respects soft tissue healing times	80%
		Inflammation marker — Diurnal pattern (24‐h behavior): Worst AM and PM; improving throughout the day	80%
		Inflammation marker — Pain is associated with cardinal signs of inflammation (i.e., swelling, redness, heat)	80%

Domain 2 – Nociplastic pain drivers
Category A: Possible nociplastic pain		Category B: Probable nociplastic pain
The pain is chronic (persists for more than 3 months)	81%	There is a history of pain hypersensitivity in the region of pain. Any one of the following	92%
		‐ Sensitivity to touch
		‐ Sensitivity to pressure
		‐ Sensitivity to movement
		‐ Sensitivity to heat or cold
The pain has a regional distribution (rather than discrete).	94%	Presence of comorbidities, any one of the following:	98%
		‐ Increased sensitivity to sound and/or light and/or odors
		‐ Sleep disturbance with frequent nocturnal awakenings
		‐ Fatigue
		‐ Cognitive problems such as difficulty to focus attention, memory disturbances, etc.
There is no evidence that nociceptive pain (a) is present or (b) if present, is entirely responsible for the pain; and	94%	Non‐consistent mechanical patterns of pain reproduction (e.g., inconsistent reproduction of pain by movement)	91%
There is no evidence that neuropathic pain (a) is present or (b) if present, is entirely responsible for the painful symptoms.	97%	Widespread pain	85%
Evoked pain hypersensitivity phenomena can be elicited clinically in the region of pain (static mechanical allodynia, or dynamic mechanical allodynia, or heat or cold allodynia, or painful after‐sensation reported following the assessment of any of the above alternatives).	100%

Domain 3 – Drivers related to neuropathic pain
Category A: Signs and symptoms resulting from damage to the peripheral somatosensory system.		Category B: Signs and symptoms resulting from damage to the central somatosensory system and/or upper motor neurons.
Pain localization is associated with a cervical spinal nerve root dermatome	98%	Evidence of myelopathy (e.g., Babinski's sign, Hoffmann's sign)	95%
Qualitative description of pain involves tingling, paresthesia/dysesthesia, burning/shooting	94%	Presence of central neurological pathology (e.g. cervical dystonia)	94%
Pain is caused by movements that stress neural tissues (mechanosensitivity of the nervous system)	95%
Presence of sensory deficit in the region of a dermatome (e.g., numbness in a root dermatome)	94%
Decreased osteotendinous reflexes (i.e.: Hyporeflexia)	83%
Motor deficits (e.g., weakness) in a plausible neuroanatomical distribution	94%
Neurological symptom modification (e.g., decrease of paresthesia during cervical movement)	93%

Abbreviations: NAG, Natural Apophyseal Glides; SNAG, Sustained Natural Apophyseal Glides; The percentages represent the proportion of participants who rated the item as “highly relevant” or “relevant.”

Table 3.

Drivers of pain and disability according to the Pain and Disability Management Model for people living with neck pain (PDDM‐NP).

Domain 4 – Comorbidity drivers
Category A: Comorbidity drivers that can be managed by clinicians		Category B: Comorbidity drivers requiring multidisciplinary management or that are outside clinician scope of practice (requiring referral)
Concomitant painful musculoskeletal condition (e.g., low back pain, knee pain, etc.)	100%	Obesity	84%
Headache of musculoskeletal origin (cervicogenic headache, tension‐type headache, etc.)	97%	Endocrine disorders (e.g., diabetes, hypothyroidism)	85%
Chronic primary pain (e.g., fibromyalgia, chronic fatigue syndrome, irritable bowel syndrome)	91%	Mental disorders (e.g.: PTSD, Somatization, Anxiety disorder, Depression, other DSM‐V diagnoses)	98%
Symptoms associated with vestibular disorders (e.g., dizziness, vertigo)	96%	Respiratory pathologies (e.g., Chronic obstructive pulmonary disease, asthma)	88%
		Cardiovascular pathologies	89%
		Systemic conditions (e.g., rheumatoid arthritis, ankylosing spondylitis)	100%
		Headache not of musculoskeletal origin (Migraine, medication overuse headache, etc.)	94%
		Diagnosis of sleep disorders (not related to the painful condition)	97%
		Substance abuse disorders	92%

Domain 5 – Cognitive‐emotional drivers
Category A: Drivers associated to cognitive‐emotional responses to pain are present and contribute to the patient's pain or disability.		Category B: Cognitive‐emotional drivers associated with maladaptive (unhealthy) behaviors that play a predominant role in the painful experience have been identified and need to be addressed (Interdisciplinary approach possible).
Self‐reported symptoms of depression	98%	Avoidance behavior (Withdrawal from painful activities, change of posture, etc.)	97%
Self‐reported stress	97%	Maladaptive coping behavior (addictive behavior around food, alcohol, etc.)	97%
High level of anxiety	97%	Social withdrawal since the presence of the condition	97%
Posttraumatic anxiety and fear	97%	Health‐seeking behaviors	91%
Pain‐related anger and frustration	98%
False beliefs in regard to pain	100%
Poor expectations of recovery	100%
Pain‐related worries	100%
Self‐efficacy (low)	100%
Kinesiophobia (fear of movement)	100%
Fear of recurrence	98%
Readiness to Change	98%
Perceived injustice	95%
Feelings of guilt	89%

Abbreviations: DSM‐V, Diagnostic And Statistical Manual Of Mental Disorders, Fifth Edition; PTSD, Posttraumatic stress disorder; The percentages represent the proportion of participants who rated the item as “highly relevant” or “relevant.”

Table 4.

Drivers of disability according to the Pain and Disability Management Model for people living with neck pain (PDDM‐NP).

Domain 6 – Environmental or lifestyle drivers, and social determinants of health
Category A: Drivers related to social environment, work environment and lifestyle factors		Category B: Social determinant of Health
Interpersonal relationships (social needs, social relationships, socialization activities, social support)	95%	Economic stability (economic resources make it possible to engage in health‐promoting behavior, etc.)	98%
Family dynamics (marital relationships or familial aspects of life such as sexuality, spousal support, or the presence of dependents).	97%	Access to education (better health literacy, contributes to psychological and physical well‐being, etc.)	92%
Work‐related characteristics (repetitive work, handling of heavy loads, work schedules, high physical demands, etc.)	100%	Access and quality of care (bias linked to skin color, difficulty getting to care facilities, difficulty paying for care, etc.).	98%
Litigation situation (legal disputes, factors related to litigation, etc.)	95%	Food insecurity (lack of food, difficulty accessing healthy food, etc.)	89%
Lifestyle factors, such as recreational/leisure activities, smoking or alcohol consumption.	98%	Neighborhood and built environment (environment that does not promote physical activity, urban planning that favors isolation, etc.).	89%
		Social and community context (influence on personal choice [smoking, diet], acceptance of individual behavior by others, inclusion, early childhood development, etc.).	95%
		Housing, basic amenities and the environment (air pollution, noise pollution, water contamination, access to facilities such as water, electricity)	91%

Note: The percentages represent the proportion of participants who rated the item as “highly relevant” or “relevant.”

Figure 2.

Pain and Disability Drivers Management Model. (A): Involves more common, modifiable elements. (B): Involves complex elements that are difficult to modify, and generally require a multi‐disciplinary approach and/or referral.

4. Discussion

The objective of this modified DELPHI study was to develop and validate the content of the Pain and Disability Drivers Management Model for patients living with neck pain (PDDM‐NP). Over 50 participants with different areas of expertize gave feedback on each component of this framework, enabling them to reach a consensus of 70 elements distributed in six domains. From the original proposal, 60 elements were validated (97%). Based on the participants' suggestions, 10 new elements were added.

4.1. What Do Our Results Show?

Early in the process, a consensus was reached for most of the elements, as 60/62 elements were validated during the first round. However, two specific elements, namely inflammatory markers and overweight, didn't reach the pre‐established consensus threshold. The inflammatory markers (e.g., pain is associated with cardinal signs of inflammation (i.e., swelling, redness, heat)) evaluated in our study were taken from the work of Kolski et al., whose aim was to develop a pain mechanism classification system for any patients seeking physiotherapy treatment [41]. As these diagnostic criteria have been identified to address all pathologies, they are not specific to neck pain. Symptoms that may describe inflammatory phenomena in other joints may be found in benign presentations of neck pain (e.g., crackling). Other symptoms, such as swelling, characteristic of inflammation affecting the knee, for example, are less characteristic of neck pain. Collectively, these elements provide an explanation for the fact that no consensus was reached regarding “Pain is described using terms such as: Swollen, stiffness, crackling, dull aching or throbbing” item. The Berlin criteria could have been proposed to our participants, but these have psychometric values slightly inferior to the Kolski criterion [33, 42].

Overweight was initially proposed in the model since it has been reported in multiple studies to be a negative prognostic factor [32]. One explanation for this is that body mass index, which is used to define overweight, is increasingly controversial (e.g., gender/ethnic disparities, no estimation of fat distribution) [43]. A solution has been suggested to associate the BMI measurement with the waist circumference measure for a better representation of risk [44]. However, the addition of this type of practice can increase the time burden of the patient examination and may also induce patient discomfort.

4.2. What's New in This Version?

In addition to adapting the content of the PDDM to patients with neck pain, two major components were modified. Firstly, we added a domain to recognize the presence of “nociplastic pain” as a potential driver of painful symptoms. In the original version of PDDM, only one mechanism of nociplastic pain was included, central sensitization, which was a part of the nervous system dysfunction domain [17]. Since its publication in 2017, the label “nociplastic pain” has been added to the pain terminology [45], and diagnostic criteria have been proposed [46, 47]. Given the abundant literature demonstrating that pain may be driven by nociplastic origin in patients with trauma‐related neck pain, as well as the presence of this type of pain in some people with nonspecific neck pain, it seemed important that this category should appear as a standalone domain in the model [48, 49].

Secondly, we have proposed that the original domain pertaining to contextual factors now considers “environmental or lifestyle factors, and social determinants of health” as the main domain driving disability. A recent study portrayed contextual drivers as “components of all therapeutic encounters and may constitute the entirety of the perceived effects of the intervention itself or be additive to the effects of interventions such as pharmacological and nonpharmacological treatment” [50]. There is emerging evidence that the social determinants of health account for a significant proportion of health outcomes [35]. They encompass a range of factors, such as education, economic stability, or access to care [51]. Although these factors are dependent on political or structural factors, they can be partially addressed by clinicians [52, 53]. Strategies such as “social needs–informed care” (i.e., modify its management according to the patient's social factors) or “social needs–targeted care” (i.e., help patients find solutions to their social problems) could be easily performed by clinicians [54]. We're confident that these changes will add value to the existing model and help rehabilitation professionals deliver person‐centered care.

4.3. Perspectives

This framework considers a range of factors that can explain the patient's experience of pain and disability. Ideally, this is performed within the context of a biopsychosocial model. However, many clinicians feel under‐trained or under‐equipped to assess and manage these drivers of pain and disability [55, 56]. This framework can support clinicians' reasoning process, increase their confidence in their perception of competence, and improve the quality of care that they provide. Previous work demonstrates the relevance of these frameworks and the importance of tools supporting psychosocial evaluation [57, 58]. Indeed, the low back pain version of the PDDM has demonstrated its relevance in guiding clinicians in the assessment of patients [21] and has demonstrated its effectiveness to optimize the management of low back pain [22]. Further steps are necessary to transfer this theoretical framework into a more clinician‐friendly format, including both measurement tools and management proposals based on the elements highlighted by the measurement tools. These steps will be carried out during a feasibility study in which we will assess the acceptability of the tool to clinicians and patients.

4.4. Strengths and Limitations

The questionnaire was developed using robust methodology to ensure that we had the most relevant elements. We established a steering committee to improve the face validity of our questionnaire. We selected clinicians with diverse experiences to ensure a variety of viewpoints. However, there are some limitations. First, a large proportion of our experts have Mechanical Diagnosis and Therapy training, which may lead to a bias in their views. This bias was limited by successive reminders to increase the number of participants from other backgrounds. Second, most of the participants were from North America, which tends to orient the model towards a North American vision of rehabilitation. Third, content validation requires the involvement of patients' perspectives, which was not the case here. This may affect the number and consensus of certain elements included in the model. These considerations must be considered when developing the clinical tool associated with this model.

5. Conclusion

Through a modified DELPHI study, we updated the PDDM model for people living with neck pain. Our expert panel reached a consensus and added new items. The resulting model contains 70 elements integrated into six drivers of pain and disability. Future research should facilitate the integration of this diagnostic framework into clinical practice.

Author Contributions

Thomas Gerard: conceptualization, methodology, formal analysis, writing–original draft, review and editing, funding acquisition. Florian Naye: conceptualization, writing–review and editing. Simon Decary: conceptualization, writing–review and editing, Pierre Langevin: conceptualization, writing–review and editing, Chad Cook: conceptualization, writing–review and editing. Yannick Tousignant‐Laflamme: conceptualization, methodology, formal analysis, writing—original draft, review and editing, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting information.

Supporting information.

Supporting information.

Supporting information.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.