|Home | About | Journals | Submit | Contact Us | Français|
Low back pain (LBP) is currently the most prevalent and costly musculoskeletal problem in modern societies. Screening instruments for the identification of prognostic factors in LBP may help to identify patients with an unfavourable outcome. In this systematic review screening instruments published between 1970 and 2007 were identified by a literature search. Nine different instruments were analysed and their different items grouped into ten structures. Finally, the predictive effectiveness of these structures was examined for the dependent variables including “work status“, “functional limitation”, and “pain“. The strongest predictors for “work status” were psychosocial and occupational structures, whereas for “functional limitation” and “pain” psychological structures were dominating. Psychological and occupational factors show a high reliability for the prognosis of patients with LBP. Screening instruments for the identification of prognostic factors in patients with LBP should include these factors as a minimum core set.
La lombalgie (LBP) est un problème important notamment sur le poste des dépenses dans nos sociétés modernes. Sur le plan évolutif, il est primordial de pouvoir dépister des facteurs pronostics négatifs de ces patients. entre 1970 et 2007, un certain nombre d’éléments permettant cette analyse ont été identifiés dans la littérature. Ces instruments d’analyse nous semblent importants de façon à déterminer la limitation fonctionnelle, les possibilités de travail et la douleur. l’élément de prévision le plus important pour la poursuite d’un emploi est psychosocial, de ce fait l’élément important pour apprécier le pronostic de limitation fonctionnelle ou de douleur est plutôt psychologique. les facteurs psychologiques et sociaux permettent de faire le pronostic des lombalgies, il s’agit d’un tronc commun des éléments de dépistage, devant inclure tous ces facteurs.
Low back pain (LBP) is the most prevalent and costly musculoskeletal problem in today’s economically advanced societies, and may lead to long-term disability combined with frequent use of health services [2, 8, 14]. In Germany, a lifetime prevalence of LBP of 80%, an annual prevalence of 60%, and a point prevalence of 30–40% lead to economic overall costs of 16–22 billion Euro p.a. . Comparable figures are stated for the UK, the Netherlands, Sweden, and the USA. Of these costs, 30% are direct costs accounting for medical treatment and 70% are due to indirect costs as loss of production. In Germany, LBP causes 4% of all loss of production .
In its natural course, nonspecific LBP is self-limiting within a few weeks whereas 3–10% of patients develop persisting LBP . Although this is only a small group of patients the socioeconomic burden significantly exceeds that for the treatment of acute LBP .
Therefore, it is of high importance to identify patients at risk for developing persisting LBP at an early stage. To detect these patients, prognostic factors of chronicity must be known. According to the biopsychosocial model the influence of different factors has to be taken into account [5, 24].
Screening instruments are needed to assess these influencing factors and to foretell the course of LBP. This review provides a survey of these instruments, analysing and classifying them according to specific aspects. Structures are evaluated following Waddell and Burton  and a compilation of all aspects, structures, and their predictive effectiveness is given.
The strategy of the literature search was based on the “method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group for spinal disorders”  and comprised three steps. First, the data-bases MEDLINE/Pubmed, ISI Web of Knowledge, and PsychINFO were searched with the terms “screening tool”, “screening instrument”, “risk assessment”, or “questionnaire and back pain”. The queries in ISI Web of Knowledge and PsychINFO were limited to publications between 1970 and 2007; the search in MEDLINE/Pubmed was not restricted. In addition, references of two key publications were included [28, 29]. Second, references of identified articles were searched with the same terms as in the first step. Third, after identification of screening instruments a final query was performed combining the name of the respective instrument with the search term “back pain”.
This literature search entailed looking into screening instruments based on studies already completed at the time of the search. Data analysis was performed in a retrospective approach.
We applied broad inclusion criteria to ensure that the spectrum of screening instruments and patients included represent the spectrum seen in routine settings. Inclusion criteria of the instruments considered were study samples of patients with unspecific acute, subacute, or chronic LBP according to the definition of the “COST B13: European guidelines for the management of low back pain” .
The search resulted in thirteen articles containing nine different instruments (Table 1):
In order to compare the effectiveness of items of the identified instruments, these items were grouped into ten structures (Table 2). Whereas some of these structures have already been established and are widely used (e.g. fear-avoidance beliefs), others have been designed by the authors by combining similar aspects. Therefore, a meta-analytic comparison of structures is not possible.
Table 2 lists all structures and their aspects. Furthermore, the predictive effectiveness of structures is given, differentiating between the dependent variables “work status“, “functional limitation”, and “pain“. “Work status” summarises ”capacity to work”, ”disability days“, ”future sick absenteeism“, ”days off-work“, ”return to work“, ”return to full-time work within three months“, ”working/not working three months after“, and ”work loss“ from different studies. “Functional limitation“ merges ”activity limitations“, ”functional limitation”, ”disability“, ”disability in activity of daily living“, ”function“, ”functional status”, and ”bed rest”. “Pain” entails “pain”, “chronicity“, and “pain severity“.
Structures were reviewed according to the following four aspects:
Evidence of structures was evaluated as strong, moderate, limited or without evidence following the scale introduced by Waddell and Burton . The evidence for the independent variable was reduced by one level if the dependent variable was listed but not examined in half of all articles (in odd numbers of articles one article less than half) or evaluated by one article only.
Twelve out of thirteen publications investigated the influence of structures on the dependent variable “work status”, seven publications studied the influence on “functional limitation”, and six on “pain”. Each of the ten structures were analysed with respect to the above-introduced four aspects and evaluated regarding its evidence.
Six articles investigated instruments containing socio-demographic factors. These instruments were ALBPSQ, ÖMPSQ, Screening Questionnaire for Predicting Outcome in Acute and Subacute Back Pain, VDPQ, and HKF-R 10. ALBPSQ, ÖMPSQ, and Screening Questionnaire for Predicting Outcome in Acute and Subacute Back Pain will be summarised to ALBPSQ due to their significant overlap of questions. Limited evidence could be demonstrated for an influence of this structure on “work status” and “pain” as dependent variables.
Five authors examined instruments comprising work characteristics and work related attitudes. These instruments were ALBPSQ and VDPQ. Strong evidence existed with regard to an influence of this structure on “work status”, with limited evidence on “functional limitation” as dependent variables.
Six articles evaluated three different instruments (ALBPSQ, CPCI, PPS) addressing physical functioning/activities of daily living. Limited evidence was noted for an influence of this structure on “work status”, with moderate evidence for “functional limitation” and “pain” as dependent variables.
Nine publications analysed five different instruments (ALBPSQ, HKF-R 10, INTERMED, PPS, VDPQ) regarding aspects of pain. Influence of this structure on all dependent variables was of limited evidence.
Four authors examined three different instruments (HKF-R 10, VDPQ, INTERMED) in consideration of medical aspects. Moderate evidence was demonstrated for influence of this structure on “work status” and limited evidence on “pain” as dependent variables.
Five studies considered two instruments (ALBPSQ and HKF-R 10) for the structure depression. Limited evidence was shown for influence on “work status” and “functional limitation”, with moderate evidence on “pain” as dependent variables.
Eight trials evaluated four instruments (ALBPSQ, HKF-R 10, INTERMED, PPS) for negative psychological states. Moderate evidence was found for “work status” and “pain”, limited evidence for “functional limitation” as dependent variables.
Nine publications investigated five instruments (ALBPSQ, CPCI, HKF-R 10, INTERMED, PPS) containing coping strategies/reaction to pain. Limited evidence could be demonstrated for influence of this structure on “work status” and “functional limitation”, with moderate evidence on “pain” as dependent variables.
Seven authors examined instruments comprising of fear-avoidance beliefs (ALBPSQ and Fear-Avoidance Beliefs Questionnaire). Strong evidence referred to influence of this structure on “work status” and moderate evidence on “functional limitation” as dependent variables.
Four articles evaluated three different instruments (ALBPSQ, CPCI, VDPQ) addressing emotional and social support system/psychosocial aspects. Moderate evidence was noted for influence of this structure on “work status”, with limited evidence for “functional limitation” and “pain” as dependent variables.
Reliability and prognostic validity of screening instruments investigated are shown in Table 1 including frequencies, percentages, p values, and coefficients if appropriate. Further details on different aspects of structures and their influence on the three dependent variables are listed in Table 2.
The pattern analysis of the ten structures (Table 2) identified predictors of the three dependent variables “work status”, “functional limitation”, and “pain” in patients with LBP. There are unspecific predictors as “physical functioning/activities of daily living” and “depression” being effective for all three dependent variables. Other predictors such as “fear-avoidance beliefs” and “emotional and social support system/psychosocial aspects” are specific for “work status” and “functional limitation”. “Fear-avoidance beliefs about work” and “perceived chance of being able to work” could be revealed as the strongest predictors of “work status” whereas “sleep” and “fear-avoidance beliefs” were found to be the strongest predictors of “functional limitation”. “Intensity, duration, and frequency of pain” and “coping strategies/reaction to pain” were specified as the strongest predictors of “pain”. When building screening instruments this pattern analysis should be taken into consideration using those predictors which are specific for the patient group.
Limitations of this review are the relatively small number of studies investigated, different patient samples regarding the duration of LBP, varying methodology, and insufficient data on patient samples (e.g. when attending a physician for the first time, LBP history). Furthermore, when grouping items into structures we were not able to include all of the items.
There is a great diversity of items in the literature describing a large number of influencing factors in patients with LBP. The ten structures developed in our review show the spectrum of the biopsychosocial model. Although there is consensus in the literature about abandoning the biomedical approach and focussing on the biopsychosocial model [5, 24], the impact of individual factors is rated differently among authors.
Pincus et al. point out the influence of psychological factors with regard to the correlation between psychological distress/depressive mood and increased risk of chronicity . This endorses the relevance of our structures “depression”, “negative psychological states”, “coping strategies”, “fear-avoidance beliefs”, and “emotional and social support system/psychosocial aspects”.
Shaw et al. list these factors as significant: “low workplace support”, “personal stress”, “shorter job tenure”, “prior episodes”, “heavier occupations with no modified duty”, “delayed reporting”, “severity of pain and functional impact”, “radicular findings”, and “extreme symptom report” . Comparing these data with our findings illustrates the limitations of comparability of both reviews. Only some items reported by Shaw et al. can be assigned to one of the ten structures of our review. Whereas we considered “low workplace support”, “personal stress”, and “heavier occupations with no modified duty” as predictors as well, “severity of pain and functional impact” and “extreme symptom report” have no prognostic influence according to our results; the allocation of the remaining factors is ambiguous.
Steenstra et al. characterised the following factors as predictors for a longer duration of sick leave in patients with LBP: “specific LBP”, “higher disability levels”, “older age”, “female gender”, “more social dysfunction and more social isolation”, “heavier work”, and “receiving higher compensation” . “More social dysfunction and more social isolation” and “heavier work” were also evaluated as predictors in our review. However, results by Steenstra et al. with respect to socio-demographic factors as being predictive did not match our findings.
In conclusion, this systematic review has found psychological and occupational factors to have the highest reliability among prognostic factors of patients with LBP. To optimise the decision-making process of physicians these factors should be included in future screening instruments for the identification of prognostic factors at an early stage. Furthermore, psychological and occupational factors should be part of a minimum core set in future prospective studies as referred to in an ongoing study of the authors’ looking into predictors of the transition from acute to chronic LBP, including the impact of different health care systems .
This project was supported by the National Research Programme NRP 53 “Musculoskeletal Health - Chronic Pain” of the Swiss National Science Foundation (Project 405340-104826).
M. Melloh, Phone: +64-3-4740999, Fax: +64-3-4747617, Email: email@example.com.
A. Elfering, Phone: +41-31-6313639, Fax: +41-31-6318212, Email: firstname.lastname@example.org.
C. Egli Presland, Phone: +41-31-6313639, Fax: +41-31-6318212, Email: hc.ebinu.stneduts@ilgenemrac.
C. Roeder, Phone: +41-31-6315940, Fax: +41-31-6315931, Email: email@example.com.
T. Barz, Phone: +49-3332-532314, Fax: +49-3332-532735, Email: firstname.lastname@example.org.
C. Rolli Salathé, Phone: +41-31-6313639, Fax: +41-31-6318212, Email: hc.ebinu.stneduts@illorailenroc.
O. Tamcan, Phone: +41-31-6315941, Fax: +41-31-6315931, Email: email@example.com.
U. Mueller, Phone: +41-31-6315942, Fax: +41-31-6315931, Email: firstname.lastname@example.org.
J. C. Theis, Phone: +64-3-4740999, Fax: +64-3-4747617, Email: email@example.com.