post-template-default,single,single-post,postid-1169,single-format-standard,elision-core-1.0.11,ajax_fade,page_not_loaded,qode-child-theme-ver-1.0.0,qode-theme-ver-4.5,wpb-js-composer js-comp-ver-6.6.0,vc_responsive
Title Image


Neck Pain

Predicting Neck Pain

  |   Research   |   17 Comments

Topic Area: Prognosis
Article: Salo P. Neck Muscle Strength and Mobility of the Cervical Spine as Predictors of Neck Pain: A prospective 6-year study. Spine 2012;37(12):1036-1040.

Article Summary:

  1. It is well documented that individuals with neck pain have significant loss of neck muscle strength and ROM …but the question is, chicken or the egg? Is it possible that those with weaker and stiffer necks are simply more likely to end up with neck pain?

The purpose of this 6-year study was to determine whether neck muscle strength or passive mobility of the cervical spine among initially pain-free working-age women could be used to predict future neck pain.
The researchers screened 220 pain-free, healthy female volunteers where isometric neck muscle strength and passive ROM of the cervical spine were measured.

After 6 years of following these patients 19% admitted to having neck pain for at least 7 days during the past year.

However, isometric neck muscle strength or passive mobility could not predict future occurrences of neck pain in healthy pain-free women. We can safely conclude that the neck weakness and loss of passive ROM comes AFTER the onset of neck pain, and not before …so the chicken came first!

Critical Appraisal

  • Was the defined representative sample of patients assembled at a common (usually early) point in the course of their disease?

The sample was indeed assembled at a common point in time.  All 220 subjects were deemed to be pain-free after answering a screening questionnaire on health status, occupation, time used for leisure time physical activity, and level of physical workload.  The sample itself was not representative of the general population as it only consisted of females that were employed at the largest companies in the city of Jyvaskyla, Finland. 

  • Was patient follow-up sufficiently long and complete?

Yes, follow-up was sufficiently long and complete.  A questionnaire package was sent to each subject 6 years after their baseline measurement.  Questions included onset of neck pain in the last year and also during the past 6 years, amount and length of neck pain, medication, health care utilization, and accidents or illnesses associated with their neck pain.  One hundred and ninety-two subjects returned the questionnaire.  In terms of length of time for follow-up, the authors admitted that 6 years was too long suggesting the results were less reliable when participants had to recall pain symptoms during such a long period.   They also indicated that the results of the study may have been stronger if the sample had been screened regularly for neck pain during the whole 6-year period.     

  • Were outcome criteria either objective or applied in a blind fashion?

While the authors did not specifically state any “blinding” methods for the researchers, the design of this study may be that “blinding” was not required since there was no treatment provided nor were measurements taken a second time.  A single tester performed the initial range and strength measurements with reliable and valid instruments used in clinical practice. Participants provided the remainder of outcome data subjectively via postal questionnaire.

  • If subgroups with different prognosis are identified, did adjustment for important prognostic factors take place?

No, as the authors were looking for isometric neck muscle strength and/or passive range of motion as predictive ability between the two groups, they used bias-corrected bootstrap CI’s to determine cut-off points for strength and mobility. Ninety-five percent CI values were used to calculate sensitivity, specificity, positive predictive value and likelihood ratios.  Differences between the groups in the area under the curve (receiver operator characteristic curves) were also evaluated. 

Additionally, the t test, x2 test, Mann-Whitney U test or Fisher Exact test (alpha level at 0.05) were used to provide statistical comparisons of baseline characteristics between those who did and did not return the survey. Statistical comparison was also made of baseline characteristics between those who did not (NPO) and did (NP) experience pain in the past year or 6 year interval from the time of initial intake.

Baseline characteristics of the subjects who did not return the questionnaire (n=28) did not differ from those who returned the questionnaire.  In both groups (pain or no pain) baseline characteristics were similar except for the physical workload.  Those without neck pain tended to participate in more medium work activities more so than those reporting neck pain at 6 months. 

  • How likely are the outcomes over time?

Of the 220 healthy female volunteers, 192 (87%) responded at 6 year follow-up via a postal survey. This is only a 13% attrition rate. Thirty-seven (19%) reported having neck pain for 7 days or greater in the past 6 years, while 155 (81%) reported neck pain for less than 7 days (or none at all) in the previous 6 years.

  • How precise are the prognostic estimated?

A 95% CI was established for isometric neck strength: .52-.56 (.41-.66), and cervical mobility: .54-.56 (44-.76). A positive predictive value for all items assessed remained between .21-.29 for all variables (strength and mobility assessment). When the area under the curve (AUC) is under .70 the test is described as a poor predictor, and if the AUC is < .6 then the test is a very poor predictor, per the University of Nebraska Medical Center1. Therefore, according to the study, both cervical mobility assessment and baseline cervical strength in healthy women are poor predictors of future neck pain.

  • Can I apply this evidence about prognosis to my patient?

The participants in the study were all healthy and were excluded if they answered positively for any neck or shoulder pain in last 6 months, had previous or current neck injuries or other disorders of the neck or shoulder area, arthritis, fibromyalgia, severe depression or mental disorder, or an active competitive sports career.   This is not representative of the typical patient seen in clinic where pain, immobility and/or loss of function often triggers a referral for physical therapy management.  Rarely do patients seek physical therapy for prophylactic treatment alone. 

Likewise, all participants were women between the ages of 20-59 which is not representative of the total population typically seen in clinic.  Furthermore, participants in the study were volunteers that were employed at the largest companies in the city of Jyvaskyla, Finland which may have also potentially biased the outcomes of the study.

The predictive indicators (isometric neck muscle strength and passive mobility) and the outcome measures (isometric neck measurement system and cervical measurement system) chosen by the authors to determine the likelihood for future neck pain was a little surprising considering the evidence linking weakness of the segmental deep neck flexors with ‘chronic’ neck pain and passive range of motion versus segmental mobility (PIVM’s).  Nevertheless, what we can take away from this study is that isometric neck muscle strength and passive ROM are not predictive of future occurrences of neck pain in pain-free working age women and perhaps other indicators need to be examined. 


  1. Tape, TG. “The Area Under An ROC Curve.” University of Nebraska Medical Center.  29 July 2012. < http://gim.unmc.edu/dxtests/roc3.htm>
  • Joseph S. | Nov 25, 2019 at 2:54 pm

    Critical appraisal of prognostic studies can be challenging but you did a really nice job with this. Two forms of bias that are typical in these types of studies are sampling bias and referral bias.

    • Janice | Nov 25, 2019 at 6:04 pm

      it is apparent that the authors failed to represent the general population in their recruitment leaving their study with elements of both sampling bias and referral bias which negatively impacts the external validity of their study. Since the author’s limited recruitment through direct advertising to the employees of the largest employers in the City of Jyvaskyla, Finland, the sample no doubt has subtle, unique characteristics which are not likely representative of the general population, leading to a false cohort (sampling bias). Likewise, this unique subgroup of the population will likely have a higher prevalence of unfavorable prognostic factors than patients in the community at large which makes generalizing the findings of this study problematic (referral bias).

  • PJ PT | Nov 25, 2019 at 5:15 pm

    Something that both you and the article touched on as a limitation was that the follow up for neck pain was only at year 6 and only included the previous year. They reference that this is the procedure that was used in a previous study, but I am not sure why it was used.
    It would seem to me that a great deal can change in the 5 years between testing and beginning to follow-up. Weakness or ROM limitations could have developed during this time period. It would have seemed that a more appropriate manner of follow up would have been a questionnaire at 1 year and 2 years. This may have yielded different results.
    What are you thoughts on why 6 years was the follow up time?

    • Janice | Nov 26, 2019 at 2:48 pm

      In terms of length of time for follow-up, you’re right 6 years was too long for patients to recall pain symptoms and no doubt influenced the reliability of the data. The authors did not indicate the reasoning behind the 6 year follow-up but did indicate that the results of the study may have been stronger if the sample had been screened regularly for neck pain (i.e. annually). I was very surprised that they chose to measure isometric muscle strength over DNF strength considering the evidence linking weakness of the segmental deep neck flexors with ‘chronic’ neck pain. I would be very interested to see whether DNF strength measured over a course of 3-5 years is an effective method of screening healthy subjects who are at risk of developing neck pain. As of yet, we do not have a reliable predictor of neck pain in health subjects. I believe the closest we have come was a systematic review by McLean et al (2010)1 who set out to answer who is at risk of developing non-traumatic neck pain? The following 7 factors were linked to the development of non-specific / non-traumatic neck pain.

      Female gender
      Older age
      High job demands
      Low social/work support
      Being an ex-smoker
      A history of low back disorders
      A history of neck disorders
      No matter how amazing our manual therapy techniques, needling techniques, muscular training exercises etc are, we as physio’s are unable to influence the top 7 risk factors affecting the recurrence rate!

      1. McLean SM, May S, Klaber-Moffett J, Macfie Sharp D, Gardiner E. Risk factors for the onset of non-specific neck pain: a systematic review. Journal of Epidemiology and Community Health. 2010; 64(7):565-572.

  • Dr. John | Nov 25, 2019 at 5:29 pm

    I agree with PJ. First, I agree that a more frequent questionnaire would have been a better way to follow-up with the patients. Also, I know that the purpose of this study was to see if DNF strength and AROM are predictors of future neck pain, but I would have loved to see the changes in DNF strength and AROM after 6 years! Of course it becomes harder to prevent drop-out with a long follow-up and requiring subjects to come back in to be tested. The authors do express in the discussion section that they would have rather sent out questionnaires more frequently to prevent decreased reliability.

  • Brenda Kurt | Nov 27, 2019 at 3:30 am

    My guess as to why 6 years was picked as the follow-up time is because that is how the previous study performed their follow-up. Why the previous study chose 6 years? I have no idea. I imagine that some of it has to do with compliance and having the ability to follow-up with participants – people move and contact information changes all of the time. I would personally like to see a longer follow-up as well, 15 to 20 years.

  • Dr. John | Dec 6, 2019 at 1:42 pm

    In this study, ROM and neck muscle strength, were not seen as predictors of neck pain. The 7 factors listed are risk factors for developing non-traumatic neck pain. The difference between these 2 groups of factors is modifiable versus non-modifiable. Carroll et al stated that identifying both is important, as with modifiable factors a suitable intervention can be matched, compared to non-modifiable factors which determines those patients with a higher risk of chronic neck pain.

    ref: Carroll LJ, Hogg-Johnson S, van der Velde G, Haldeman S, Holm LW, Carragee EJ, et al. Course and prognostic factors for neck pain in the general population: results of the Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders. Spine. 2008;33(4 Suppl):S75-82. Epub 2008/02/07.

    • Janice | Dec 10, 2019 at 1:50 pm

      It was concluded in the appraisal that ROM and decreased muscle strength come after the neck pain. These factors are modifiable, and from this the appropriate interventions can be applied. When it comes to the non-modifiable factors, our best approach would be patient education. Unfortunately, once neck pain does develop, and patients are classified, the corresponding intervention can be applied, and here is where we as clinicians can put our skills to use.

      Childs JD, Fritz JM, Piva SR, Whitman JM. Proposal of a Classification System for Patients with Neck Pain. J Ortho Sports Phys Therapy. 2004;34:686-696.

  • Burda | Jan 1, 2020 at 9:21 am

    I look forward to further information from you on this topic

  • Melvin Loaps | Jan 8, 2020 at 12:54 am

    This is very helpful to my practice.
    Thank you.

  • John | Apr 7, 2020 at 2:29 am

    Hi Janice,
    Great Job!
    Thanks for the great information.

  • buford | Apr 13, 2020 at 9:25 pm

    Amazing blog. Such a great share. Thanks a lot for sharing these interesting ways to control our breath to become happy.
    I am glad to visit and read your blog. Keep it up.

  • Eric | May 3, 2020 at 11:45 pm

    Greetings! Quite helpful guidance on this post! It really is the small changes that make the biggest changes.Thanks a lot for sharing!

  • Zurek | Jun 9, 2020 at 12:17 am

    Awesome post thanks for sharing a lot of good information

  • Justin | Jun 16, 2020 at 8:44 pm

    Very nice article.

    I really enjoyed going through all the great information you’ve provided in this great article.

    Keep up the great work.

    Thank you

  • physical therapy thousand oaks ca | Nov 24, 2020 at 2:58 am

    Physiotherapists in most jurisdictions, such as the Australian states & most of the United States, and of course all of Canada enjoy professional autonomy, with the ability to act as primary care providers and to determine — and be responsible for — a patient’s management plan. Physiotherapists in some other jurisdictions work primarily upon referral from other professionals (typically medical practitioners). Another great post!

Post A Comment