Best Pract Res Clin Rheumatol. 2013 (Oct); 27 (5): 601–612 ~ FULL TEXT
Iben Axén, Charlotte Leboeuf-Yde
Unit of Intervention & Implementation Research,
Institute of Environmental Medicine,
Nobels väg 13, S-171 77 Stockholm, Sweden
Low back pain is not a self-limiting problem, but rather a recurrent and sometimes persistent disorder. To understand the course over time, detailed investigation, preferably using repeated measurements over extended periods of time, is needed. New knowledge concerning short-term trajectories indicates that the low back pain 'episode' is short lived, at least in the primary care setting, with most patients improving. Nevertheless, in the long term, low back pain often runs a persistent course with around two-thirds of patients estimated to be in pain after 12 months. Some individuals never have low back pain, but most have it on and off or persistently. Thus, the low back pain 'condition' is usually a lifelong experience. However, subgroups of patients with different back pain trajectories have been identified and linked to clinical parameters. Further investigation is warranted to understand causality, treatment effect and prognostic factors and to study the possible association of trajectories with pathologies.
From the FULL TEXT Article:
Until recently, low back pain (LBP) was believed to be a self-limiting condition, much like the
common cold. The European guidelines for the management of acute LBP state that 90% of patients will
recover within 6 weeks. 
However, research in the past two decades has shown that the majority of back pain sufferers
experience episodes of the problem; LBP is actually a recurrent condition.  Thus, LBP resembles a long-term condition such as asthma rather than a self-limiting condition such as the common cold. This means that we need to look at LBP (and perhaps all types of spinal pain) as a lifelong process, perhaps with different causes and modifying factors as life goes on, but always present as an underlying ‘trait’.
This shift of paradigm directs the focus of attention away from LBP seen as a single entity to the LBP condition regarded as a chain of LBP episodes , that is, with the focus on the course of LBP.  This new paradigm comes with a hope of being able to classify nonspecific LBP into clinically meaningful subgroups.  Such subgroups may offer new insights into causal mechanisms, prognostic factors and effective treatment strategies. In the past, most studies related to LBP as though it has an ‘end point’, measuring for example, pain intensity, return to work, or recovery at a rather arbitrarily chosen point in time with questionnaire surveys. Because patients with LBP experience ups and downs of varying intensity and duration, such end points are likely to capture LBP at different phases of the condition, making comparison between people and studies potentially meaningless. These outcomes should instead be viewed in the light of the long-term development of LBP (see also chapter 2 of this edition). However, little is known about the detailed course, and trajectory, of LBP.
Methods of data collection for LBP trajectories
In order to study a pain trajectory, frequent measurements of pain, disability or any other outcome
over a fairly long time are needed, as measures may vary considerably with time in a fluctuating
condition. Until recently, this was not really possible using postal surveys, as this would be both
difficult and expensive. Instead, realistically, frequent data collection could be achieved only by using paper-based diaries, which have been shown to produce data of questionable validity.  With modern technology, it is now possible to use Web-based diary systems to facilitate repeated and frequent data collection. However, this requires the respondents to have Internet access and to be disciplined in using it also at frequent intervals. This is likely to result in biased study samples.
Frequent e-mail messages is another method that is cheap and capable of collecting ‘fresh’
momentary data much like electronic diaries, but – again – requires easy and frequent use of the
Internet plus the discipline suiting the rhythm of questions. In other words, study participants must be able and ready to answer the questions as they are sent out, frequently in some cases but less frequently in others.
In some recent prospective studies, text messages and mobile phones have been used to collect
frequent data. Questions are sent on scheduled days and times to the included respondents, who
answer with a reply text message in a fully automated Web-based system, SMS Track.  This system has the advantages of being cheap  and user-friendly , as most people in the modern world carry their phone with them at all times and thus can respond at any time. Further, it has been shown to be capable of yielding valid data.  However, the questions asked are restricted by the size of the text message, and must be short, requiring only a short answer. Examples of how such questions can be worded are provided in Table 1.
Answers are immediately recorded in a data file, accessible to the researcher online at any time.
Depending on the question asked, the respondents may, for instance, record their pain on a scale from
0 to 10, as it is right now, the past 24 h, the past week and so on. The choice of measure and the
frequency of data collection must of course be matched to minimise memory bias and decay. For
example, the optimal recall period for pain would not be the same as that for sick leave.
However, one could be concerned that frequent and repeated measurements may tire respondents.
Nevertheless, the response rates in these studies are high, so far between 63%  and 82%.  Moreover, in one study, patients from secondary care answering two text message questions weekly for 52 weeks said that they would not have minded a third question.  Another worry might be that the intense focus on the respondent’s conditions may actually influence the examined outcome and that the constant reminding may lead to mood problems. This was tested in a study of patients with rheumatic pain and found not to be the case. 
In the near future, the use of tailor-made smart phone applications will most certainly combine the
advantages of the diary with those of the mobile phone, as more sophisticated electronic equipment
becomes common property within all social classes and as Internet access becomes increasingly
available and affordable. Frequent data collection therefore offers many new research opportunities.
In addition, clinicians can make use of this method. The course of a condition can be illustrated on
an individual level and used in the clinical decision making as well as to sensitise patients to the long-term
condition. Some examples of some individual patterns are shown in Figsutes 3–6 later in the paper.
The LBP condition versus the LBP episode
It is important to conceptualise LBP as representing two different aspects of a single disorder: the
long-term condition of being prone to having LBP versus the acute episode, relapse or exacerbation,
when the pain appears, reappears or worsens. If we, again, compare LBP to asthma, we can imagine that the individual LBP episodes may be prevented and treated just as the individual asthma attacks
often can be handled. However, the true underlying cause of asthma is not the inhalation of pollen, and
similarly one can assume that the causes of the LBP condition are not the same as those of the LBP
episodes. The fact that this concept has not been clearly understood in the past may well explain our
poor understanding of causality, treatment choice and prognostic factors. We therefore argue that
these two concepts must be studied separately.
As discussed in the previous chapter by Dunn et al., LBP is a lifelong problem for most.  Ideally, to fully understand the LBP condition, lifelong studies should be undertaken, starting with the very first onset of pain to map out the development over time in terms of characteristics, frequency and duration. Clearly, this is not possible. We therefore have to make do with less lengthy observations to identify different particular patterns. It is also difficult to capture the very first onset of LBP in children and adolescents, for which reason we would have to make do with studying the course pattern in people who already have developed LBP.
Ideally, the study of LBP episodes should capture people who start a new episode after a period of
being LBP free ; however, this may be difficult. The closest we might get to this is studies of people who consult for their problem, hoping that many of these have just started a new episode. Nevertheless, experience shows that patients often decide to consult later in the course of an episode and not at the very onset. In two observational studies [13, 14] of patients consulting for LBP in the primary care setting, around 40% stated that their pain started within the previous week, around 15% stated that it was of 2 weeks’ duration and the remaining 45% had had pain for longer than a fortnight.
In order to obtain a detailed picture of the trajectories of LBP, frequent data collection points are
necessary. Only then will the pattern of relapses and remissions be possible to ascertain.
Below follows a description of what is presently known about the long-term and short-term trajectories
of LBP in adults.
The overall course of LBP – the LBP condition
In our opinion, ideally, the course of LBP should be measured at least monthly over a minimum
period of 1 year. Lemeunier et al. conducted a systematic review of longitudinal studies of LBP in the
general population  and set the requirements at a minimum of 100 patients followed up with a minimum of three surveys. They found eight original studies that were quite diverse with regard to the definition of LBP, populations under study, length of follow-up (between 1 and 28 years), number of follow-ups and outcomes. Unfortunately, only one study  fulfilled our criteria for frequent data collection. Despite these differences, a synthesis of the results revealed that people without LBP largely continue to be LBP free and those who have LBP continue to report LBP, indicating very clearly that LBP continues its course over time, with very little change. LBP is indeed a persistent condition. However, the different trajectories that people experience over their lifetime have not been mapped out nor is the clinical importance of various lifetime trajectories known.
The short-term course of LBP
Studies over 1 year
Itz et al.  conducted a systematic review of prospective studies of patients with LBP of <3 months’ duration from the primary care setting  and concluded that spontaneous recovery is unlikely, as 65% of patients still experience pain after 12 months. In their review, none of the included studies fulfilled our criteria for frequent data collection, and thus the conclusions were based on fewer measurements.
A summary of the findings from the literature in which the short-term, hence more detailed, course
of LBP was studied is presented in Table 2, using information from the individual studies described in
Tamcan et al.  collected weekly data in the general population using diaries for 1 year on pain intensity, medication use, health-care use and social and work limitations in people from the general population who had reported LBP at an earlier survey. They identified three different persistent courses and one fluctuating course.
Monthly surveys of pain intensity for a year were collected by Dunn et al.  in patients from the primary care setting, all of whom consulted for LBP, resulting in similar findings, two sub-trajectories of persistent pain and one of fluctuating pain. However, also a group of ‘recoverers’ was identified. Contrary to the two studies above, a study by Jensen et al.  found LBP to be persistent over 1 year in patients who had consulted the secondary care setting (Figure 1). In their study, weekly data were collected using SMS Track on the number of pain days each week. A further analysis of these data by Leboeuf-Yde et al. showed that the vast majority (80%) never experienced a non-episode during this period of 1 year, a non-episode being defined as at least 4 weeks in a row without any LBP.  This is hardly surprising, as this group would be likely to consist of people with a more chronic condition than people in the general population or in the primary care setting.
Leboeuf-Yde et al., in another study, analysed data on 50-year olds from the general population  and showed that there are three distinct and equally common pain trajectories, namely those who during 1 year never have LBP, those who have it episodically and those who have persistent LBP. In addition, these data were collected with frequent text messaging and this study also included subjects
without LBP at baseline.
Studies over less than 1 year
Menezes et al. conducted a review and meta-analysis of 33 clinical studies with follow-up times
between 3 and 12 months , which concluded that among acute LBP sufferers, the majority had recovered by 12 weeks, whereas among patients with persistent LBP, less than half had recovered at that time. None of the included studies used frequent data collection and were therefore not able to illustrate the detailed course over time.
In a study by Eklund et al. using weekly text message data over 6 months in patients with LBP from
primary care , the subgroups detected in the general population described above  were replicated, although the proportions were somewhat different. As this was a study of patients consulting for LBP, no pain-free individuals were found. However, as expected, the clinical population from the primary care setting contains more people with persistent pain than the general population and fewer than secondary care. 
As in the review by Menezes , early improvement was noted also by Kongsted and Leboeuf-Yde , who collected 18 weekly text messages from patients in the primary care setting. This study revealed a quick reduction in LBP reporting, which slowed down towards the end of the study (Figure 2). This initial favourable development was also confirmed by Axén et al. in a 6-month study from primary care using weekly text message data.  On a group level, improvement from the LBP episode was seen mainly in the early course. LBP in this study was reported as number of days each week with bothersome LBP. Although improvement was the common development, four subgroups with different after-treatment trajectories were identified.
Group 1 was the ‘typical’ trajectory: definite improvement
was noted in the first 5–6weeks (Figure 3).
Groups 2 and 3 responded faster and slower, respectively (Figures 4 and 5).
Finally, Group 4 showed a rather stable trajectory (Figure 6) with mild, persistent pain.
Figs. 3–6 below show some individual examples of each of these types of trajectories (Fig. 6).
In addition, Kongsted and Leboeuf-Yde  noted a number of sub-trajectories in relation to early course and later course over the duration of the study of patients in primary care, which lasted for 18 weeks. More than 50% of the subjects could be categorised as experiencing early improvement followed by full recovery or staying somewhat improved. An additional 30% could be classified as ‘fluctuating’.
Trajectories of the LBP episode
None of the studies reported above included obvious inception cohorts in terms of new episodes of
LBP. Jensen et al.  had an inclusion criterion of at least 2 months and Tamcan et al.  studied patients who previously reported LBP in a cross-sectional survey. Leboeuf-Yde et al.  studied people
from the general population (with or without LBP at baseline) and in the studies by Dunn et al.  Kongsted et al.  and Axén et al./Eklund et al. [22, 23] (who used data from the same study sample),
all the included subjects had consulted for LBP. Probably, in some of these primary care populations,
patients experiencing their first LBP episode may be present, but it is much more likely that
the bulk of these patients had experienced LBP previously. However, no such distinction was made in
any of the studies. Therefore, presently there seems to be no information available on the trajectories of
first-time LBP episodes.
Summary of trajectories of LBP
Studies of trajectories of LBP using frequent data collection methods have identified various subgroups.
As can be seen in Table 2, several methods of classification were used, and LBP was studied both
in the clinical and in the general population. Still, the identified subgroups have similar features,
persistent and episodic patterns being identified by all research groups. The sizes of these subgroups
vary, as would be expected in different populations.
Pain trajectories related to clinical findings
It is all very well to detect and describe how LBP runs a different course in different groups of
people. However, trajectories will only become clinically relevant when they can be used to investigate
the cause and prognosis of a condition.
Does it matter to which subgroup people belong?
Do different trajectories have different aetiologies?
Should they be treated differently?
Can the prognosis be altered?
A first step would be to see if different trajectories have specific clinical characteristics. To date,
only four studies appear to have coupled LBP trajectories with clinical data [5, 10, 16, 23] and one study investigated whether periods of LBP were correlated with amount of leisure-time physical activity. 
Dunn et al.  subgrouped patients on the basis of pain intensity and found that changes in disability showed similar patterns to changes in pain. They also noted some differences between groups on psychological variables such as depression. The patients in the ‘persistent mild’ category
showed large reductions in depressive symptoms following primary care consultation, compared to
the ‘severe chronic’ patients who had high depression scores with no improvement. In addition, patients
belonging to the ‘fluctuating’ pain group failed to show improvements in depression scores. However, it was not possible to discern the cause and the effect between the psychological findings and the pain trajectory.
Tamcan et al.  identified their subgroups according to pain characteristics, medication use, health-care use and social and work limitations and noted that they coincided with different trajectories. In addition, differences between these groups on the basis of self-rated health and demographic characteristics were assessed. A logical patternwas found: older patients and those with poorer quality of life were most commonly found in the ‘severe persistent’ and ‘moderate persistent’ groups
compared to the ‘fluctuating’ and ‘mild persistent’ groups.
Kongsted and Leboeuf-Yde linked validated diagnostic subgroups using pain trajectories in patients
with LBP in the primary care setting.  In this study, patients with disc pain had more days with pain and very rarely (only 5%) belonged to the ‘recovered’ pain course. Patients with pain of a muscular origin, on the other hand, reported fewest days with pain and a third of these patients belonged to the ‘recovered’ trajectory.
Axén et al. , after having identified pain trajectories, tested whether these were statistically associated with some clinical background variables. Although only few variables were available, it was possible to identify significant and meaningful differences between the four trajectory subgroups and these clinical characteristics. For example, those with the quickest recovery started off being in the worst shape. Those with the slowest recovery were also the oldest with the most persistent pain at baseline.
Finally, Jespersen et al.  coupled the pain pattern over 1 year with leisure-time physical activity in a cohort of employed subjects with or without LBP at baseline. No correlation between the two was detected, indicating that leisure-time physical activity is not protective of LBP, contrary to common belief.
In summary, the results from these five studies clearly indicate that this research area deserves further exploration. In other words, different types of trajectories seem to identify clinically meaningful
subpopulations. Finally, perhaps it will be possible to add a vital piece to the puzzle of LBP!
Despite differences in the LBP patient populations, outcomes, data collection methods and frequency
of measurement, researchers who have synthesised the recent research results have presented a
congruent picture of the LBP trajectory. It is no more to be considered a benign and self-limiting condition
but one of longer duration. Therefore, it is relevant to study both its natural and clinical course.
In the general population, when observed over 1 year with frequent data collection, there are three
large groups in relation to troublesome LBP:
those without LBP,
those who experience it on and off and
those who have it most of the time.
The bottom line is that people with LBP are likely to experience either a persistent or an episodic trajectory. Definite recovery with no recurrences does not appear to be a common outcome, although it is not known how these patterns develop over a lifetime. People with LBP will not necessarily seek care, but a person who consults a primary care clinician for an episode of LBP is likely to feel better fairly quickly. However, the post-treatment trajectories can vary, as there are several subgroups. Clinical findings have been coupled with distinct trajectories in a meaningful way, indicating that these trajectories can be used to determine clinically relevant subgroups. However, when patients with LBP have come as far as consulting in secondary care, they are likely to be persistently troubled over the ensuing 12 months.
Research agenda |
We suggest that in clinical studies, the trajectory should be used as an outcome more often. Thus, rather than just using a single end point at a single point in time, change in the LBP trajectory, measured with repeated measurements, seems a more relevant measure of effect in a disorder that does not, in fact, have a clinical end point.
The exploration of clinically relevant subgroups based on trajectories needs further attention, also using frequent data collection methods. As suggested in recent studies, these subgroups may represent different diagnoses and psychological profiles. Therefore, they may provide a key to the selection of different treatments.
Finally, the identification of the persistent and episodic nature of the LBP condition invites the idea of secondary and tertiary prevention. Research should be directed towards effective measures to minimise the occurrence, length and intensity of episodes, as well as to identify suitable methods to lessen the impacts of a persistent problem rather than attempting to cure the condition.
Practice points |
Information about LBP trajectories may inform clinical practice: The LBP condition should not be viewed as curable. We therefore suggest that clinicians should view and communicate information about episodes within the context of a longer-term pain trajectory, to provide patients with a realistic view of the problem. Thus, issues such as effective short-term treatment strategies, pain management and activity maintenance as well as secondary and tertiary prevention should be high on the clinical agenda. ‘Management rather than cure’ might be a helpful catch phrase, similar to the well-known recommendation of ‘don’t worry – keep active’.
van Tulder M, Becker A, Bekkering T, Breen A, Carter T, Gil del Real MT.
European Guidelines for the Management of
Acute Nonspecific Low Back Pain in Primary Care
European Spine Journal 2006 (Mar); 15 Suppl 2: S169–191
van den Hoogen HJ, Koes BW, van Eijk JT, Bouter LM, Deville W.
On the course of low back pain in general practice: a one year follow up study.
Annals of the Rheumatic Diseases 1998 Jan;57(1):13–9.
Kea Dunn, editor.
Best practice & research in clinical rheumatology chapter on low back pain
across the lifecourse 2013.
von Korrf M.
Studying the natural history of back pain.
Spine (Phila Pa 1976) 1994;19((18 suppl.):2041S–6S.
Dunn KM, Jordan K, Croft PR.
Characterizing the Course of Low Back Pain:
A Latent Class Analysis
Am J Epidemiol. 2006 (Apr 15); 163 (8): 754–761
Stone AA, Shiffman S, Schwartz JE, Broderick JE, Hufford MR.
Patient compliance with paper and electronic diaries.
Controlled Clinical Trials 2003 Apr;24(2):182–99.
SMS-Track. SMS-Track questionnaire 1.1.3.
Johansen B, Wedderkopp N.
Comparison between data obtained through real-time data capture by SMS
and a retrospective telephone interview.
Chiropractic & Osteopathy 2010 May 26;18(1):10.
Axen I, Bodin L, Bergstrom G, Halasz L, Lange F, Lovgren PW, et al.
The use of weekly text messaging over 6 months was a feasible method
for monitoring the clinical course of low back pain in patients
seeking chiropractic care.
Journal of Clinical Epidemiology 2012 Apr;65(4):454–61.
Kongsted A, Leboeuf-Yde C.
The Nordic Back Pain Subpopulation Program: Individual Patterns of
Low Back Pain Established by Means of Text Messaging:
A Longitudinal Pilot Study
Chiropractic & Osteopathy 2009 (Nov 17); 17: 11
Broderick JE, Vikingstad G.
Frequent assessment of negative symptoms does not induce depressed mood.
Journal of Clinical Psychology in Medical Settings 2008 Dec;15(4):296–300.
de Vet HC, Heymans MW, Dunn KM, Pope DP, van der Beek AJ, Macfarlane GJ, et al.
Episodes of low back pain: a proposal for uniform definitions to be used in research.
Spine (Phila Pa 1976) 2002 Nov 1;27(21):2409–16.
Leboeuf-Yde C, Rosenbaum A, Axen I, Lovgren PW, Jorgensen K, Halasz L, et al.
The Nordic Subpopulation Research Programme
Prediction of Treatment Outcome
in Patients With Low Back Pain Treated By Chiropractors --
Does the Psychological Profile Matter?
Chiropractic & Osteopathy 2009 (Dec 30); 17: 14
Leboeuf-Yde C, Axen I, Jones JJ, Rosenbaum A, Lovgren PW, Halasz L, et al.
The Nordic Back Pain Subpopulation Program:
The Long-term Outcome Pattern
in Patients With Low Back Pain Treated by Chiropractors in Sweden
J Manipulative Physiol Ther. 2005 (Sep); 27 (7): 472–478
Lemeunier N, Leboeuf-Yde C, Gagey O.
The natural course of low back pain: a systematic critical literature review.
Chiropractic & Manual Therapies 2012;20(1):33.
Tamcan O, Mannion AF, Eisenring C, Horisberger B, Elfering A, Muller U.
The course of chronic and recurrent low back pain in the general population.
Pain 2010 Sep;150(3):451–7.
Itz CJ, Geurts JW, van Kleef M, Nelemans P.
Clinical course of non-specific low back pain: a systematic review
of prospective cohort studies set in primary care.
European Journal of Pain (London, England) 2013 Jan;17(1):5–15.
Jensen RK, Leboeuf-Yde C, Wedderkopp N, Sorensen JS, Manniche C.
Rest versus exercise as treatment for patients with low back pain
and Modic changes. A randomised controlled clinical trial.
BMC Medicine 2012 Feb 29;10(1):22.
Leboeuf-Yde C, Jensen RK, Axen I.
Absence of low back pain in patients followed weekly over one year
with automated text messages.
Chiropractic & Manual Therapies 2012 Mar 29;20(1):9.
Leboeuf-Yde C, Lemeunier N, Wedderkopp N, Kjaer P.
Evidence-based classification of low back pain in the general population:
one-year data collected with SMS Track.
Chiropractic & Manual Therapies 2013;21(30).
da C Menezes Costa L, Maher CG, Hancock MJ, McAuley JH, Herbert RD, Costa LO.
The prognosis of acute and persistent low-back pain: a meta-analysis.
Canadian Medical Association Journal 2012 Aug 7;184(11):E613–24.
Eklund A, Jensen I, Lohela-Karlsson M, Leboeuf-Yde C, Axén I.
Absence of low back pain in patients from primary care followed with
weekly automated text messages for 6 months.
BMC musculoskeletal disorders 2013. Submitted.
Axen I, Bodin L, Bergstrom G, Halasz L, Lange F, Lovgren PW, et al.
Clustering patients on the basis of their individual course of low back pain
over a six month period.
BMC Musculoskeletal Disorders 2011 May 17;12(1):99.
Jespersen T, Jörgensen M, Hansen J, Sogaard K.
The relationship between low back pain and leisure time physical activity
in a working population of cleaners-a study with weekly follow-ups for 1 year.
BMC Musculoskeletal Disorders 2012;13:28.
Kongsted A, Leboeuf-Yde C.
The Nordic Back Pain Subpopulation Program:
Can Low Back Pain Patterns
Be Predicted From the First Consultation With a Chiropractor?
A Longitudinal Pilot Study
Chiropractic & Osteopathy 2010 (Apr 29); 18: 8
Return to the LOW BACK PAIN Section