Although neck pain can be recurrent, severe, disabling and is a socioeconomic burden on society [1, 2], it is most commonly simple and transient. It is estimated that one third of adults will experience neck pain over the course of one year [1].The balance of evidence suggests that psychosocial distress has a negative impact on the course of neck pain [3-6], possibly to a greater extent than clinical or physical variables [7]. Chronic neck pain is associated with poor psychological health, including cognitive distress, anxiety and depressed mood, a potential loss of employment and associated reduction of income, increased stress, physical limitations/pain, a reduced self-concept, changes to their social life/status and issues relating to medical adherence [7, 8]. There are significant socioeconomic implications. The most common presentations seen by psychologists after an Motor Vehicle Accident (MVA) are pain, depression and post-traumatic stress [8]. Helping people/patients change behavior may be an important factor when addressing lifestyle modifications needed for neck pain prevention and management [8]. This overview (review of review) addresses a diverse group of seemingly eclectic interventions that include interventions that may help people understand, independently self-manage, potentially limit the impact of injury on their neck pain/disability or prevent potential injury. Some interventions are intended to manage psychological distress related to neck pain and others such as patient education strategies, use of orthosis (i.e. collar), ergonomic workplace interventions, and preventions strategies, affect change in some aspect of the individual or their environment to aide in the management of neck pain.

A number of psychological treatments (i.e. cognitive-behavioral approach, interpersonal behavioral activation, supportive counselling, etc.), psychosocial interventions (i.e. multidisciplinary care, psychologist as advisor to other health professionals), and mind-body wellness approaches (i.e. yoga) exist [8]. Psychological interventions delivered by a psychologist, psychosocial intervention delivered by non-psychologists health professionals and mind body wellness delivered by a certified instructor are options considered by physicians and other health practitioners when psychological distress arises in the presence of neck pain [8].

Patient education is a foundational component of health care and an essential part of communication between patient and the healthcare provider [9, 10]. The World Health Organization (WHO) defines therapeutic patient education as education helping patients acquire or maintain the skills they need to manage their life with a disease in the best possible way [10].

Medical devices are generally defined as products (article, instrument, apparatus or component) used for medical or therapeutic aims including: (a) the diagnosis, treatment, mitigation or prevention of a disease, disorder or abnormal physical state, or its symptoms, and (b) restoring, correcting or modifying a body function or the body. Class 1 medical devices (orthosis) such as specific pillows, collar use or taping may be recommended.

Ergonomic workplace interventions may include physical redesign changes at the workplace (i.e. ergonomic workstation modification), organizational system changes at the workplace (i.e. job rotation, production system changes) or individual worker changes (education, exercise, use of orthotics) [11]. The aim of ergonomic change is often prevention. Prevention strategies are defined as preventing the onset of symptoms (primary prevention) and reducing the recurrence rate of signs and symptoms (secondary prevention) [12].

A number of these interventions may help people understand, independently self-manage, potentially limit the impact of injury on their neck pain and related disability or prevent potential injury.

Psychological treatments, psychosocial often multidisciplinary interventions and mind-body wellness approaches are designed to influence psychological process that underpin and contribute to pain, distress and disability. Interventions like cognitive behavioral therapy and neuroscience education potentially target cognitive-emotional sensitization (and descending facilitation) in patients with chronic neck pain. They aim to help individuals adjust to the reality of their chronic condition, adhere to a treatment regimen that will manage and reduce symptoms, and improve their quality of life despite their condition [13]. Some psychosocial interventions may target single elements such as stress or anxiety. Other psychoeducation provides information to better understand and cope with pain. Still other behavioral interventions may target compliance and adherence to a regimen through a program of self-monitoring and rewards. Cognitive interventions help individuals analyze their beliefs, thoughts and emotions as well as challenge irrational negative thoughts. Not unlike other chronic diseases, chronic neck pain can give rise to secondary co-morbid diagnoses requiring specific psychological treatments concurrently. Finally, various mind-body approaches like yoga breathing may reduce levels of anxiety, depression and stress and increase optimism [14].

Patient education is an iterative process: an educational diagnosis is made, a tailored education programme is established, group or individualised patient education is planned and provided, acquired skills are assessed, and the education programme is revised [15]. It is anticipated that through patient education strategies knowledge is transmitted, skills are acquired, and abilities are potentially maintained. Educational strategies for patients with neck pain may help patients achieve independence in care and self-management. However, in the context of neck pain, the optimal mode of delivery, and message to be delivered, is yet unknown. While education is a relatively low-cost intervention, in the absence of sound evidence of effectiveness even this is difficult to justify.

Orthosis maintain the neck in alignment or restricts movement minimizing symptom onset or recurrence. Similarly, ergonomic interventions may incorporate workplace physical redesign, organizational system changes or individual worker changes with the aim of reducing or preventing a worker’s signs and symptoms related to neck injury.

One overarching foundation to most of these interventions is that many are ‘change interventions’ [16]. Change interventions help people alter their maladaptive and unwanted behaviors and are especially useful in dealing with lifestyle modifications for disease or injury prevention and management [16]. Understanding the patient’s readiness to change, related barriers to change, and relapses may be important to not only improving patient satisfaction but also to producing and maintain health outcomes.

The number of clinical trials on neck pain is continually increasing [12] and an ongoing collaborative, interdisciplinary and internationally represented effort at evidence synthesis is warranted. The purpose of this overview was to systematically review existing reviews and establish evidence-based recommendations on psychological interventions, patient education strategies, orthosis use, ergonomic change, and prevention strategies for acute to chronic neck pain including WAD with and without cervicogenic headache or radiculopathy compared to primarily a control group at immediate post treatment to long-term follow-up. Our primary outcomes of interest were pain, function, disability, work related function, patient satisfaction, global perceived effect and quality of life. Our secondary outcomes include psychological measures such as depressive symptoms.

Selection criteria are noted in Table 1.

Our systematic overview process used independent assessment by at least two members of our review team and included comprehensive computerized search strategies including MEDLINE, EMBASE, CINAHL, ILC, CENTRAL and LILACS from January 2000 to August 2010. Note that two separate searches were performed - one for treatment and one for harms. The protocol for this review was not registered. Further, we undertook a grey literature search by identifying on going systematic reviews near completion such as Cochrane Reviews up to 2012, by contacting our expert panel and by systematically checking reference lists of primary studies to minimize the risk of missing relevant reviews and trials.

We utilized the following triage rules (set a priori) to guide decisions and group treatment reviewed:

1) Type of treatment was used to group reviews by treatment category

2) Within a treatment category we grouped review data with respect to comparator treatments.

3) We prioritized the highest quality reviews (low risk of bias) based on the rules below, per grouping.

a. If there were few reviews, we included them all due to the paucity of information.

b. If there were several reviews reporting on the same treatment and comparator, we selected the reviews with the lowest risk of bias high Assessment of Multiple Systematic Reviews (AMSTAR) scores for inclusion, according to the approach recommended by Whitlock [11] by considering the following:

i. Year of publication. We selected the most current reviews when the data was similar across reviews and when the studies from the older review were also included in the most recent review. Further we ensured consistency among review conclusions before eliminating older reviews. Inconsistency and discordance were highlighted and reasons for these differences were discussed;

ii.AMSTAR - risk of bias. We prioritized reviews with a low risk of bias for inclusion in our overview. Reviews that scored 8 or higher on the 11-point AMSTAR scale were considered at low risk of bias; 5 to 7 moderate risk of bias; and 4 or less a high risk of bias (See Table 2). These various reviews were further summarized in the summary of findings table with the goal of providing definitive summaries to inform clinical practice. Inconsistency and discordance were highlighted and discussed across reviews;

iii. Effect size estimates: We selected the effect size as the primary summary measure. Within a grouping for treatment and comparator, we selected a review that represented the best estimate of the effect size (or a meta-analytic effect). In cases where there was a large discordance between reviews, we reported our re-analysis of the meta-analysis using the individual studies included in the reviews. Additional data on magnitude of effect such as number-needed-to-treat (NNT), standard mean difference (SMD) and weighted mean difference (WMD) were extracted when available. We further considered the clinical importance of the treatment effects using the following guiding principles: evidence on the minimal detectable change and the minimal clinically important difference for the outcome, with a change from baseline of > 15% representing the minimal clinical important difference when not otherwise determined, the magnitude of the treatment effect (represented by WMD, SMD, NNT, absolute benefit, treatment advantage), the evidence for a dose response gradient, and evidence on the duration of effect [18].

4) Strength of Evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach: We used the same prioritized representative systematic review to judge an overall GRADE for the strength of the body of evidence for each type of treatment. The selected reviews may have already reported a GRADE table. The GRADE approach on the quality of evidence from primary trials considers information on design randomized controlled trials (RCT), information on timing of outcomes (immediately post treatment to long term follow-up); risk of bias or equivalent methodological quality criteria reported in the review i.e. PEDro (Physiotherapy Evidence Database) [19,20]; imprecision (sample size); inconsistency; indirectness and reporting bias. We applied similar principles to evaluate those systematic reviews that did not undertake an evaluation of strength of evidence.

Data extraction was performed by one reviewer and checked by a second; disagreements were resolved by consensus. We systematically extracted data from selected systematic reviews and developed evidence tables. Factors extracted from the original reviews included the following: 1) descriptive features of the original review such as authors, publication year, source disorder, symptom duration, intervention, comparator (placebo, no care, usual care, other treatment), the search period, the original authors of primary studies; 2) methodological issues of original review such as search period, AMSTAR score, quality ranking system, evidence statement and a final strength of the body of evidence rating using the GRADE tool, and 3) data on benefits and risks, including the direction and magnitude of the effect of primary outcomes with duration of follow-up, reported harms, quality ranking system, evidence statement and final GRADE. Harm information was summarized qualitatively from both the treatment reviews that reported any adverse events and directly from the harms reviews.

We used a systematic review methodological quality assessment tool (AMSTAR) [21].

Qualitative assessment using the GRADE approach [22, 23] and recommendation presentation was used. When the original review did not report using GRADE methodology, an assessment of the summary of finding estimate for GRADE was based on data extracted in the systematic review. Once reviews were deemed relevant and of low risk of bias, we reported trial findings by “quality of evidence” using the GRADE approach and by “treatment category” for the Summary of Findings Tables (Tables 3 and 4). Those intervention categories achieving moderate to high quality rating on GRADE were recorded in the recommendation table (Table 5). Associated primary outcome effect sizes and their related magnitude of effect [SMD, WMD, relative risk (RR)] were translated to clinically relevant terms (WMD, NNT) to aid knowledge users in judging their clinical importance.

We found no trials that met this standard.

We found no trials that met this standard.

We found 1 trial (moderate GRADE, 268 participants, 2 publications, Manca et al. 2007 [55], Klaber-Moffet et al. 2005 [56]) on psychosocial intervention delivered by a PT based on cognitive–behavioral principles. It encouraged a return to normal daily activities as soon as possible through self-management compared to usual physiotherapy. At long–term follow-up, there was a moderate effect size for neck pain and disability SMD 0.30 [0.06, 0.54]; Mean Difference 1.90 [0.39, 3.41] favoured the usual physiotherapy treatment over the brief intervention group. This brief intervention provided only slightly less health benefit on average in the long-term as measured by Patient-specific quality-adjusted life years at long term follow-up.

For intensive relaxation training (progressive relaxations, functional relaxation, autogenic training, systematic desensitization) delivered by a physiotherapist for chronic non-specific neck pain, there was no difference (moderate GRADE, 1 trial, 258 participants, Viljanen et al. 2003 [57]) in neck/arm pain relief or improved function when compared to a control or exercise at long-term follow-up.

For office workers with neck pain, relaxation training including progressive relaxation, autogenic training, functional relaxation, systematic desensitization showed no significant difference in pain levels or disability over no treatment (moderate GRADE, 1 trial, 258 participants, Viljanen et al. 2003 [57]). Likewise, one further trial (stress management versus no intervention, low GRADE, 192 participants, Horneij et al. 2001 [58]) showed no significant difference for pain or sick leave at long-term follow-up.

For computer and office workers with neck pain, we note no difference between two-component work-place interventions (work style group) including stimulated workplace adjustment, improved body posture, breaks, and coping behaviors to high work demands over no intervention on pain severity at long-term follow-up (moderate GRADE, 1 trial, 2 arms - 210 and 209 participants, Bernaards et al. 2007 [59]). The second arm (work style + physical activity group) included an increase in moderate to heavy physical activities. Two further trials included work ergonomic education and stress coping stills (moderate GRADE, 1 trial 2 arms, 192 participants, Andersen et al. 2011 [31]; low GRADE, 1 trial - 2 arms, 42 participants, Andersen et al. 2008 [60]) compared to various exercise interventions and found either the 2 or 12 minute exercise was favoured (Andersen et al. 2011 [31]) or no difference when compared to general fitness or neck/shoulder exercise (Andersen et al. 2008 [60]).

Four trials [61-64] (moderate GRADE, 1 meta-analysis [65], 1546 participants) suggests a number of physical environment interventions do not prevent neck pain in the short-term when compared to no intervention for non-sick listed neck pain free (university, kitchen, computer or call centre) workers while three other physical environment interventions (low GRADE, 1 meta-analysis [65], 3 trials [42, 64, 66], 295 participants) did not prevent neck pain in the long-term when compared to a control. Sick leave was also no different from a control [64]. For workers with neck pain, we note three further physical environment modifications (low GRADE, three trials Voerman et al. 2007, Fostervold et al. 2006, Hedge et al. 1999 [67-69], 137 participants) suggesting no evidence of benefit in the long-term for computer or office workers with chronic neck pain when compared to another workplace intervention. We refer the reader to Table 4 for descriptions of the modifications.

We suggest two categories of exercise 1) upper extremity stretching and strengthening or 2) general exercise may not reduce pain intensity or pain prevalence in workers. Upper extremity stretching and endurance training (moderate GRADE, one trial, 265 participants, Viljanen et al. 2003 [57]) for chronic neck pain and function at immediate post treatment, short- and long-term follow-up and a teacher guided combined exercise approach (Feldenkrais Intervention, low GRADE, one trial, 38 participants Lundblad 1999 [70]) of stabilization of the low back and pelvis, posture awareness, ergonomic training, and strength, coordination, endurance, flexibility/smoothness and rhythm exercises when compared to no intervention or a wait list control for pain reduction with in chronic neck pain at short-term follow-up showed no evidence of benefit. General fitness training (i.e. biking) (low GRADE, 1 trial, 30 participants, Andersen et al. 2008 [60]) or group exercise that combined extensibility and coordination exercises with cardiovascular training (low GRADE one trial, 44 participants, Takala et al. 1994 [71]) for pain reduction immediately post treatment in patients with neck pain of unspecified duration did was no differ from a control group.

When a soft collar is used following acute WAD and is compared to general physiotherapy, the research showed moderate evidence of no benefit (5 trials, 637 participants, Giebel et al. 1997, Rosenfeld 2000, Schnabel et al. 2004, Vassiliou et al. 2006, Bonk et al. 2000 [72-76]). These trials concluded that there was a greater reduction of pain and increased global effect at short- and intermediate-term follow-up but not at long-term follow-up with physiotherapy treatment versus continuous use of a collar. Similarly, there is moderate evidence of no benefit (3 trials, 641 participants, McKinney 1989, Mealy et al. 1986, Kongsted et al. 2007 [77-79]) when comparing the use of a soft collar in combination with education to self-mobilize when compared to general physiotherapy [77-79]. Following an acute WAD, the evidence shows that general physiotherapy is better at reducing short-term pain, and increasing short-term cervical ROM than the combined use of a collar and education to self-mobilize [78, 79]. However, there is no significant difference when comparing the effects of these two interventions on long -term pain or function [77].

For a self – management educational strategy (moderate GRADE, 1 trial; 180 participants Ylinen et. al. 2003 [33]), utilizing education on aerobic exercise plus stretching compared to specific neck, trunk, lower and upper extremity exercises or specific neck endurance and strengthening exercises was not beneficial for chronic non-specific neck pain and disability at long-term follow-up. Additionally, we note moderate evidence of no benefit (1 trial, 218 participants, Chiu et al. 2005 [80]) for a self-management strategy on neck care versus intensive neck exercises for acute neck pain and disability at short- and long-term follow-up.

Cognitive-behavioral therapy delivered by clinical psychologist for chronic neck and shoulder pain was not different from cognitive-behavioral therapy delivered by psychologist functioning as a coach to other health professionals (low GRADE, 1 trial, 70 participants, Jensen 2005 [81]) for intermediate-term pain relief or disability and long-term pain relief or sick leave. By the intermediate-term, the costs of delivering the intervention favour the coached health professionals.

Two trials assessed two different psychosocial interventions conducted by a physiotherapist. For cognitive behavioral components integrated with PT delivered by a physiotherapist in chronic WAD, there was no benefit (low GRADE, 1 trial, 33 participants, Soderlund 2001 [82]) over PT alone for pain relief, disability and activity of daily living at short-term follow-up. There were four phases to the cognitive-behavioral therapy: Learning of basic physical and psychological skills, application and generalisation of basic skills in everyday activities derived from functional behavioral analysis and a maintenance phase. One further trial (low GRADE, 1 trial, 139 participants, Vonk 2009 [83]) assessing a Behavior Graded Activity Program showed no difference in pain relief, function, global perceived effect and self-efficacy from a conventional exercise program at short-, intermediate- and long-term follow-up but required less treatment. Finally, the Neck Book provided by a physician from an occupational healthcare network intended to allay unrealistic fears of patient and promote activity despite pain (Very low GRADE, 1 trial, 181 analysed/522 randomized participants, Derebery 2009 [84]) noted a great loss to follow-up (n=181 randomized/552 analysed) and make conclusions difficult but demonstrated poor short-term and intermediate- term benefits in pain/disability. One further trial of very low GRADE noted stress management (169 randomized/282 analysed participants, Horneij et al. 2001 [58]) for mechanical neck disorder (duration NR) was not beneficial when compared to a non-intervention group for pain relief at long-term follow-up. The profession of the practitioner delivering the intervention was not reported.

For an alternative mind-body approach – Gestalt therapy 'philosophy of life' training, we found evidence of no benefit (very low GRADE, 1 trial, 87 participants, Ventegodt 2004 [85]) when compared to no treatment for pain, daily functioning, sick leave and global quality of life at short-term follow-up.

For computer workers from large office organizations with work related neck symptoms, computer stimulated breaks showed no evidence (low GRADE, 1 trial, 219 analysed/268 randomized participants, van den Heuvel 2003 [86]) of altering pain intensity/frequency or sick leave but did alter productivity and perceived recovery from complaints.

For healthy pain free workers, doing strengthening exercise (very low grade, 1 trial, 19 participants, Hamberg-van Reenen et al. 2009 [87]) did not protect against new onset musculoskeletal discomfort in the short-term.

Three trials suggest no significant change in pain prevalence [88], pain intensity [89, 90] or sick leave [90] when various combinations of four component workplace interventions were compared to no intervention at intermediate-term (low GRADE, 1 trial, 2 arms – 54, 57 participants; Ketola 2002 [89]; very low GRADE, 1 trial, 76 participants, Kamwendo et al. 1991 [90]) or long-term follow-up (low GRADE, 1 trial, 3 arms - 601, 599, 629 participants; Morken et al. 2002 [88]).

There is very low GRADE evidence of no difference (1 trial, 196 participants, Gennis 1996 [91]) when comparing the use of a soft collar following an acute WAD versus rest for short term pain relief or global perceived effect. The research also showed low GRADE evidence of no difference (1 trail, 211 participants, Borchgrevink et al. 1998 [53]) when comparing the use of a collar versus advice to “act as usual” for immediate, short-term or intermediate-term pain relief, global perceived effect, improving ROM, or time on sick leave. Also there was not difference in pain at the short- and intermediate-term follow-up if a collar was worn 2 days versus 7 days (1 trials, 70 participants, Dehner et al., 2006 [92]).

Following an acute WAD, there is very low evidence of no benefit (1 trial, 108 participants, Crawford et al. 2004 [93]) when comparing the combined the use of a collar for 3 weeks followed by a self-mobilization exercise regime versus a control group. This trial found that there was no benefit of collar use on pain, ROM, activities of daily living or return to working in the short-term, intermediate-term or long-term. Advice to self-mobilize was administered by providing participants with advice sheets. Furthermore, there is very low evidence of no difference (1 trial, 135 participants, Pennie 1990 [94]) when comparing the initial use of a soft collar followed by home exercises and late physiotherapy when compared to manual therapy for the treatment of acute WAD. Collar in combination of exercise and physiotherapy showed no difference in reduction of pain and global perceived effect in a short-term and intermediate-term follow-up.

When compared to active physiotherapy, there is low level evidence of no difference (2 trials; one with a 2 year follow-up, 286 participants, Kuijper et al. 2009, Persson et al. 1997 & Persson et al. 2001 [95-97]) for semi-hard and rigid collar use in those with cervical radiculopathy [95-97]. Evidence shows that there is no difference between these two treatments when comparing short-term intermediate-term and long-term pain, function, global perceived effect and long-term disability [95-97].

When compared to surgery, there is also low level evidence of no difference (2 trials; one with a 2 year follow-up, 149 participants, Kadanka et al. 2000, Persson et al. 1997, Persson et al. 2001 [96-98]) for semi-hard and rigid collar use in those with cervical radiculopathy. The evidence shows that there is no significant difference in short- and long-term pain, global perceived effect, and disability as well as intermediate- and long-term function.

Low GRADE evidence of no benefit was shown for the following 11 educational interventions:

1) Advice on activation versus another treatment [52] or home exercise [99] for mechanical neck pain at short-term [52, 99] and long-term follow-up and function [99];

2) Neck school which included instruction for exercise, relaxation and self-care versus no treatment for mechanical neck pain and function at short- and intermediate-term follow-up and knowledge [90];

3) Lecture on neck pain and recommendations on exercises versus multimodal treatment of exercise and relaxation training or home exercise for chronic neck pain for short- and long-term pain and function [99];

4) Education or control booklet versus no treatment for acute or subacute neck pain for short-term pain and intermediate-term pain and disability [84];

5) Education on return to normal activity using media, manual/book, video tape interview and role playing versus physiotherapy for subacute or chronic neck pain for intermediate- and long-term pain, function and quality of life [56];

6) Physician advice versus naprapathy for acute neck pain for short-term pain [100];

7) Whiplash pamphlet versus generic information sheet for acute WAD for short-term pain, function, recovery and global perceived effect [101];

8) Advice to act-as-usual and education on WAD versus immobilization in a collar [53, 102] or manual therapy and exercise [102] for acute WAD for short- and intermediate-term pain and global perceived effect [53] and long-term pain, function and quality of life [102];

9) One hour verbal advice plus a hand-out of key points versus an educational pamphlet for acute WAD for short- and intermediate-term pain and disability [103];

10) Mobilization advice versus rest for acute WAD for short-term pain and ROM [78];

11) Early or delayed standard treatment which included written material advising for rest followed by active movement versus early or delayed active treatment for acute WAD for intermediate- and long-term pain and ROM [74].

We note low GRADE evidence of no benefit for pain and stress-coping skills education from the following 2 trials: 1) education on psychological skills plus physiotherapy versus physiotherapy for subacute WAD for short and intermediate term pain and disability [82]; and 2) education on stress management versus individual training programme or no intervention for mechanical neck pain and disability at short- and long-term follow-up [58].

Also, low GRADE evidence from 2 trials suggests workplace education on pain and stress-coping skills was not favoured. First, health counselling versus specific training or general fitness training for chronic neck pain at short-term follow-up [60]; Second, education on physical exercise, advice to stay active, diet, smoking, alcohol use, stress management and workplace ergonomics versus 2 minute or 12 minute exercise program for chronic neck pain at short-term follow-up [31].

We also note that a variety of educational self-management strategies appear not beneficial (Low GRADE, 5 trials) as follows. 1) Self-care booklet versus massage plus advice on stretching, body awareness and water intake for chronic neck pain, function and quality of life at short-, intermediate- and long-term follow-up [104]; 2) Health promotion versus strengthening arm and shoulder exercises plus static neck exercises or non-specific exercises for neck pain, work ability, sick leave and discomfort at long term follow-up [32]; 3) Anatomy and stretching education versus self-stretching plus neck ROM exercises for non-specific neck pain and pressure pain threshold at short-term follow-up [105]; 4) Self-management program which included education on exercise versus multimodal physiotherapy for chronic pain from WAD at short-term follow-up [106]; 5) Instruction in home exercises versus supervised training rehabilitation for acute neck pain, disability and sick leave at short-term follow-up [107].

There continues to be a lack of high-quality evidence to inform recommendations on the use of psychological interventions, orthotics- collar, pillows, kinesio taping, or oral splint use, patient educational strategies, ergonomic workplace interventions and prevention for neck pain. We found 5 trials meeting the moderate quality evidence threshold and reporting evidence of benefit. For psychological interventions specific to chronic neck pain, there is one mind-body interventions (Dantian Qigong exercise) that assists 1 in 4 to 6 participants a small amount in pain reduction and 1 in 5 to 8 participants in function in the short-term when compared to a wait list control. Of 19 trials on patient education strategies, there is one that supports the use of an educational video given in an Emergency Room for acute WAD and helps 1 in 23 participants reduce pain a small amount in both the short- and long-term. For secondary prevention of acute to chronic myofascial neck pain in white collar workers, small daily amounts (2 or 12 minutes) of progressive resistance training with elastic tubing over 10 weeks reduced pain a small amount in 1 or 4 participant immediately post this treatment period. There is a lack of moderate or strong evidence in other treatment categories upon which to make recommendations. There is emerging and promising evidence from many small low GRADE trials that need verification. In these trials, it is the imprecision (sample size) and indirectness rather than low risk of bias assessment that limits us from making recommendations.

There was more data on moderate evidence of no benefit at short- and long-term follow-up: 2 psychosocial interventions delivered by a PT; 4 workplace interventions; 2 soft collar interventions and 2 self-management educational strategies. Psychosocial interventions by PTs not effective in long-term chronic neck pain management were solution finding using cognitive behavioral principles or intensive relaxation training. Workplace interventions that were ineffective in the long-term pain prevalence were mental health education (intensive relaxation training ), ergonomic education plus mental health education (work style or work style plus physical activity training; weekly email on this topic), individual worker change (upper extremity stretching and strengthening endurance training - 3 time per week for 8 weeks) and physical environment changes (ergonomic training plus workplace adjustment in university workers, kitchen workers, computer workers, support call center workers at short-term follow-up). The use of a soft collar or soft collar plus advice was consistently inferior to usual physiotherapy care at long- and intermediate term follow-up respectively. Finally, self-management educational strategies were less effective compared to exercise for reducing pain and disability in the long term.

Lastly, the trials we found rarely report adverse event data. There were no reviews reporting adverse events in this category of treatments. This is consistent with a review of cognitive behavioral therapy for low back pain patients reported that side effects and adverse events related to this intervention are unknown [108]. Interestingly a prospective study with a two year follow-up of the effect of CBT on patients with chronic low back pain suggested that CBT may indirectly decrease adverse events since participants decreased antidepressant drug use [109].

We came to realize through our reviewing process that expanding on three facets of our search method could lead to broader overall completeness and applicability of results from our overview. First, for MVA WAD prevention the inclusion of biomechanical studies and cohort trials might better inform overview results. For example, one recent review found a reduction of MVA related WAD claims and subjective complaints of 21 to 49% when energy-absorbing seats aimed at decreasing occupant loads and accelerations were used. Additionally, this same review showed the use of active head restrains that provide early head support to minimize head and neck motions showed evidence of reducing WAD claims and subjective complaints between 43 to 75%. Epidemiological outcomes were changed drastically for WAD prevention [110]. Another example is neck injury rate from use of helmet in biking accidents gleaned only from cohort trials [111, 112]. Clearly, studies addressing anti-whiplash systems may reduce WAD injury through improved prevention strategies especially if they are adapted by industry, the government and the clinic setting. Missing key messages like this is a limitation of this review’s search strategy.

Second, we found an absence of psychological interventions performed by a psychologist. A broader search strategy that is not limited to neck pain would enhance our understanding of the treatment for depression or post-traumatic stress by a clinical psychologist and may be applied to those with comorbid conditions of depression and neck pain [113]. Third, are therapeutic educational strategies dependent on the specific disorder or other educational dimensions? The inclusion of educational strategies for chronic musculoskeletal pain may inform us about a greater breadth of educational options not specifically studied in randomized trials for neck pain. Should we be evaluating education for chronic diseases? For example, we excluded a systematic review informing us about the compelling effect of a neuroscience educational strategy addressing a patients understanding of neurophysiological or the neurobiology effects of pain on chronic musculoskeletal pain [112].

Overall in this overview, there was incomplete and limited data. Our overview therefore has limited impact on informing or guiding clinical practice or recommendations. While many of the interventions discussed under moderate GRADE in this overview have a small magnitude of effect but lasting effect on the neck pain and to a lesser extent on function, there are many interventions (low GRADE) that require a second larger trial conducted in an independent setting to impact on our confidence in the estimate of effect. Future research needs a clear navigational strategy mapped out so we may be able to address the effects of some of these promising interventions and decrease the use of limited resources on imprudent trials.

Reporting ‘adherence to the intervention’ was poorly or not at all monitored in most trials. Since the results in different studies might be influenced by the therapists’ and patients’ adherence to treatment regime, the adherence should be better investigated in future studies to be able to analyse adherence as a covariate in effect analyses. In addition, the validity of the therapists’ skills in delivering e.g. psychological or cognitive behavioral interventions should be measured and reported in future studies. These measures would strengthen the clinical applicability and interpretability of the study’s results.

We agree with Andersen et al. [114] and Leyshou and colleagues [115] that the outcomes - productivity (cost) and safety - need to be consistently evaluated. Ergonomic interventions at the work place may alter productivity even when patient comfort (pain and function) may not change. On the one hand, if an intervention is cheap, easy to apply, easy to comply with and safe, we may consider treating even if a small difference is achieved for the short-term. For example there is evidence for some interventions that some types of specialized pillows act to provide relief from morning pain [31-32, 47, 116]. Though this effect is not a long-term effect, it may be clinically meaningful to the patient. On the other hand, if the treatment is expensive and carries further risks, we may not choose to treat.

Finally, in order to use the interventions discussed in this paper effectively with patient populations, clinicians need to determine the patient’s readiness to change and choosing the appropriate treatment interventions accordingly. We note, two trials [82]; Ferrari 2005 [101] from one review [30] on educational interventions for WAD consider readiness to change in their trials intervention. The rationale for choosing the optimal treatment for a specific patient must involve consideration of the mechanism of action of the treatment, the patient’s readiness to change and how these match with the clinical indications for treatment, as well as evidence that therapeutic objectives can be achieved. By determining the patient’s readiness to change, based on the transtheoretical model, we may be able to identify which patients are best suited to an intervention and thus better translate the evidence to the individual patient. The Stages of Change Model – pre-contemplation, contemplation, preparation, action, maintenance and relapse [16] - can help us understand success and failure rates in aiding patients with achieving change and the role of clinicians in improving patient outcomes. The physical and psychological impairments associated with the neck disorder need to be specifically targeted; consider the patient’s (and their environment’s) readiness to change; and incorporate plausible implementation strategies. Patient education, orthosis use, ergonomic change, and prevention interventions may need to be underpinned by sound adult learning principles, application of stages of change model to a practice setting, and cognitive behavioral theory. Therefore, to establish optimal treatment for a specific patient it is clear that we need conceptually sound and theory driven research in this area.

Two perspectives - AMSTAR rating of reviews assisted in triaging the evidence and GRADE was either extracted (most often from Cervical Overview Group (COG) or Cochrane reviews where a relatively consistent GRADE application process was used) or assessed by going back to the original articles to establish the quality of the body of evidence. For low GRADE evidence, further research is very likely to have a significant impact on our confidence in the estimate of effect and may change the estimate while with very low GRADE evidence we remain extremely uncertain about the estimate.

We had several strengths regarding our approach to summarizing this literature. We used a comprehensive, librarian-assisted search and multiple databases. We accessed grey literature through early submitted data for Cochrane Reviews. But we did not use new primary trials published since review updates. For example, one large (n =3851) trial (the MINT trial – Managing Injury of the Neck Trial) [117] may have an important impact on the direction of education recommendations. Two independent reviewers determined article relevance, assessed the AMSTAR methodological and verified data extraction. Our review used unique methodology, unlike Andersen [114] and colleagues’ overview methodology where overlap of original publications is detailed followed by reporting of the aim and main conclusion for each intervention, we created summary of finding tables categorized by the quality of the body of evidence (GRADE). We returned to the description of the original papers, the related GRADE statement and their reported data to be able to extract evidence on the magnitude, NNT, and duration of the effect with related side effects / harm. Consensus was used both by the internal data abstraction team and an interdisciplinary external panel consensus to validate the GRADE of evidence and recommendations. Thus, one potential source of bias is that there are several individuals on this ICON Review panel that were authors of the review papers discussed in this paper. However, these specific authors did not participate in the selection of papers or data extraction. The potential of this particular bias was reduced by using a stringent protocol for this review (as documented in the methods section).

The main findings in this paper were consistent with those found in previous reviews. In a 2010 overview, psychosocial interventions, patient education, orthotic use, ergonomic change and prevention strategies received minimal clinical directives [12, 118]. Psychological interventions were briefly mentioned. Our overview has expanded on previous reviews findings but continues to find a dearth of information on psychological interventions by psychologists. One would need to expand this review’s scope for a complete understanding on psychological interventions. For acute WAD evidence on use of an educational video on activation in the emergency department and avoidance of collar use has been well documented in the past; we have not added new information and are in agreement with others [12, 119]. For work interventions targeting the individual worker such as strengthening, the addition of new evidence (Kay et al.) has shifted the evidence from conflicting [120] or limited [121] to moderate GRADE evidence that strengthening of the cervical and scapulothoracic region aides in either preventing or reducing the recurrence of neck and shoulder pain for workers in both the short and long term. We are in agreement with Andersen (2011) [31] and Driessen 2010 [65] that no strong evidence exists for work place interventions and that there is low quality evidence in support the use of a physical ergonomic intervention (arm board support) to reduce neck pain in the long-term. For chronic neck pain or WAD, there is limited evidence emerging in favour but more against the use of psychosocial interventions. Systematically planning the future research agendas for Neck Care is needed to avoid exploring ineffective treatment pathways. We agree that targeted research of high quality is needed in work place interventions for chronic neck pain and WAD (Verhagen 2007 [122], Andersen 2011 [31]). Prevention strategies from biomechanical studies and cohort trials that showed evidence of benefit included energy absorbing seats and active head rests [24] need to be included in future overview updates.