Psycho-social factors and coping strategies as predictors of chronic evolution and quality of life in patients with low back pain: A prospective study

1. Introduction 

With a cumulative incidence as high as 70%, low back pain (LBP) represents an important factor of disabling chronic pain and low quality of life in the adult population, and a devastating problem of Public Health because of its tremendous medical and social costs. Time to evolution plays a major role: acute LBP improves in 70–90% of the cases (Andersson, 1999) and appears relatively easy to deal with, whereas chronic LBP (CLBP) is at the same time the most important source of disability and expenses, and the most difficult to treat. The 4–7% of CLBP patients who do not return to work within six months account for 75% of the medical and social costs (Maetzel and Li, 2002). So preventing a chronic evolution of LBP appears to be of great importance. This needs to identify when and why a LBP becomes chronic. Temporal definitions vary in the literature. Some authors, wishing to put the stress on the importance of an early treatment, are recommending to look at LBP as chronic as soon as one month and a half (Atlas, 1996). But in most of recent reviews and consensus reports acute LBP is defined by a pain duration within three months, and CLBP by a pain duration over three months (Bigos et al., 1994; Duquesnoy, 1994; Clinical guidelines for the management of acute low back pain, 1996, Guidelines for the management of back-injured employees, 1993).

From a clinical point of view, LBP is a complex, multifactorial process which depends on constitutional, somatic, psychological and environmental factors as well, so that the classical medical approach which deals mainly with somatic factors often fails to prevent a chronic evolution. With regard to psychological and sociological factors, many works demonstrated that they are of minor importance as far as acute LBP is concerned, but play a major role in CLBP (Truchon and Fillion, 2000). In a recent and exhaustive review of the literature, Linton concluded that “psychosocial variables generally have more impact than biomedical or biomechanical factors on back pain disability, and are clearly linked to the transition from acute to chronic pain disability (Level A evidence)” (Linton, 2000). Some works put the stress on sociological factors, showing that job dissatisfaction, low incomes and leisure are found associated with high levels of CLBP (Pope, 1989, Symonds et al., 1996, Vällfors, 1985). Among psychological factors, anxiety and depression have been for a long time the best identified of these factors (Clauw et al., 1999, Epping-Jordan et al., 1998, Fisher and Johnston, 1998, Kessler et al., 1996, Pincus et al., 2002, Polatin et al., 1993, Schermelleh-Engel et al., 1997). Cognitive factors such as fear-avoidance beliefs and passive coping are also related to pain and disability with a Level A evidence (Linton, 2000). Self-reported feelings of disability and irrational and/or negative beliefs about pain such as kinesiophobia and fear avoidance have been found associated with a chronic evolution (Fritz et al., 2001, Gatchel et al., 1995, George et al., 2003). Distraction, catastrophizing and praying/hoping are linked with negative aspects of CLPB (pain intensity, disability and poor emotional state) whereas cognitive restructuration and pain control are associated with positive evolution of CLBP (Dozois et al., 1996, Haythornthwaite et al., 1998, Kröner-Herwig et al., 1996, Lewandowski, 2004, Lin and Ward, 1996, McCracken et al., 1998, Riley et al., 1999, Robinson et al., 1997, Tuttle et al., 1991). However most of these studies are retrospective and/or cross-sectional, which prevents from analyzing precisely which psychodynamic process occurs when LBP becomes chronic. Moreover, asking retrospectively subjects who are feeling pain for a long time prevents them to provide reliable data about their psychological states and affects before the onset of pain (Koleck, 2001).

In the context of prospective studies, catastrophizing and fear-avoidance beliefs were found to be the most powerful predictors of pain and disability (Burton et al., 1995, Hasenbring et al., 1994, Klenerman et al., 1995, Linton and Hallden, 1998, Picavet et al., 2002). Passive coping (Potter and Jones, 1992), praying/hoping (Burton et al., 1995) and self-perception about the future development of pain and disability (Hazard et al., 1996, Linton and Hallden, 1998) were also found predictors of disability at a one-year-outcome.

But understanding how these variables interact with each others remain difficult, because of confounding factors, and most of these studies consider only pain criteria: chronic pain, onset of a new episode, or functional ones: disability, return to work, although quality of life and relationships between functional and emotional factors remain less documented. So prospective studies are needed to clarify the role of psycho-social factors in chronic both functional and emotional evolution of LBP.

The aim of the present study was to identify in an exploratory approach how patients deal with acute LBP, and how these reactions would expose to, or protect from, a chronic evolution and low quality of life.

2. Materials and methods 

2.1. Study sample 

Ninety-nine patients were enrolled consecutively and prospectively in the study when consulting a general practitioner (GP) for an acute LBP, and/or having stopped working or performing domestic activity because of pain within three months.

Exclusion criteria were:

•recurrent LBP or evidence of a previous LBP episode within two years,

•LBP not related to common, degenerative origin: inflammatory diseases, cancer, discal or spinal infection, Paget’s disease, osteoporosis…,

•recent spinal trauma1 (less than one week before),

•severe psychiatric disorders with hospitalization.

Ninety-nine patients fulfilling these criteria were recruted. They were 48 men and 51 women, (mean age=42.2). They were interviewed by the first author between 10 and 90 days after the onset of LBP. Table 1 provides data about these patients, and about the initial LBP episode. One year later (Table 2), 54 patients related that they had improved and returned to work within 48.3 days on average after the first interview. Six other patients had a relapse during the year of follow-up, but were free of pain at one year. So these 60 patients were retrospectively classified as improved patients (IP) at the date of follow-up. Thirty patients were still suffering from LBP at this date, and thus were classified as chronic LBP patients (CLBPP). Last, nine patients were lost for follow-up.

Table 1. Demographical, social and medical data on patients in sample and initial pain episode

	Age from 18 to 60 year old; mean=42.2; SD=11.5
	Gender: male 48; female 51
	Marital status: married or similar: 71%; bachelors: 20%; divorced: 8%; widowed: 1%
	Number of children: 0: 21%, 1: 19%, 2: 34%, 3: 19%, 4 and more: 7%
	Obesity or excessive weight: 24%
	Income: low: 12%, middle: 56%, high: 23%
	History of spinal impairments: 31%, including:
	Scoliosis 13%, Scheuermann disease 2%, listhesis 2%, Spinal trauma: 14.1%
	Previous LBP episode over 2 years: 53.3%
	Sciatica: 20.2%, including 3% with surgery
	Data on the initial LBP episode (T1):
	Evidence of discal pathology on CT-scan: 4%; facet joint pathology: 2%
	Drug treatment: analgesics: 96.9%; anti-inflammatory drugs: 75.7%
	LBP due to work-related accidents: 64.6%

Table 2. Evolution of the initial episode (T2)

	Group	Evolution	Men	Women	Total
	IP	Improvement	19	35	54 (60%)
		Improvement after relapse	4	2	6 (6.7%)
	CLBPP	Chronic pain	16	7	23 (25.5%)
		Worsening chronic pain	6	1	7 (7.8%)
		Lost for follow-up	3	6	9

3. Methods 

3.1. Inclusion data 

At the inclusion time (Tl), the aims of the study were explained by the practitioner. Patients received written information for consent, and had a face-to-face interview to record demographic and medical data. Then questionaires were administered in order to assess: (A) general psychological functioning and (B) transactional processes. All questionaires were validated, either previously, or for this study, using either principal components analysis (PCA) or structural confirmatory analysis.

(A)General Psychological Functioning:

•Trait form of State-Trait Anxiety Inventory (STAI, Spielberger et al., 1983) to assess general trends towards anxiety (validated by Bruchon-Schweitzer and Paulhan, 1993).

•Center for Epidemiologic Studies Depression-Scale (CES-D, Radloff, 1977), in a revised version developed by our group to assess trait-depression. A PCA performed on the answers of the 99 present patients revealed one factor (18 items, α of Cronbach=0.89).

•Locus of Control Scale (LCS, Lumpkin, 1985) to assess the influence a person thinks he/she may have on what happens to him/her. This scale was validated by our group in a population of 347 adults including the 99 present patients: a PCA revealed one factor (6 items, α=0.60).

•one question assessing job satisfaction of the patient.

(B)Transactional processes:

•State form of State-Trait Anxiety Inventory (STAI, Spielberger et al., 1983) to assess the state of anxiety during the interview (validated by Bruchon-Schweitzer and Paulhan, 1993).

•Center for Epidemiologic Studies Depression-Scale (CES-D) to assess the state-depression during the interview (Radloff, 1977, Fuhrer and Rouillon, 1989). A PCA performed on the answers of a sample of 125 LBP patients including the 99 present revealed one factor (18 items, α=0.83).

•Cancer Locus of Control Scale (Pruyn et al., 1988), in a revised version we developed for chronic pain patients. A PCA performed on the answers of the 125 LBP patients revealed 3 factors: perceived control on pain course (CLCS-1, six items, α=0.65), irrational beliefs about pain (CLCS-2, four items, α=0.64), and internal causal attribution of pain onset (CLCS-3, five items, α=0.67).

•Perceived Social Support Scale (PSSS): created for this study, this scale was derived from Social Support Questionnaire (Sarason et al., 1983), to assess patients’opinion on the social support they may hope for. A PCA performed on the answers of the 125 LBP patients revealed two factors: quality of social support (SSQ-1, four items, α=0.86) and availability of social support (SSQ-2, four items, α=0.63).

•Coping Strategies Questionnaire (CSQ, Rosenstiel and Keefe, 1983) to assess how patients are coping with pain. Structural confirmatory analyses performed by our group in a population of 330 chronic pain sufferers proved the pertinence of a five-factor model: distraction (CSQ-1, five items), catastrophizing (CSQ-2, five items), reinterpreting pain sensations (CSQ-3, four items), ignoring pain sensations (CSQ-4, four items), praying (CSQ-5, three items). This structure converged with the one of Robinson et al. (1997) and Riley et al. (1999) but the coping self-statement factor was suppressed because its internal consistency was too weak.

•Visual Analog Scale (VAS) to assess pain intensity during the interview.

3.2. One-year follow-up data 

•One year later, medical reports assessed how long patients have stopped working because of LBP, how many times they have consulted their general practitioner (GP) for LBP and GP’s opinion on LBP evolution.2

•State form of STAI and CES-D were used again to assess the states of anxiety and depression.

•Two questionnaires assessed quality of life:

–Medical Outcome Study short form General Health Survey (Stewart et al., 1988), a shortened form of the SF36 (Ware and Sherbourne, 1992), assessed functional limitations (MOS-1, 9 items, α=0.90), feeling of well-being (MOS-2, 4 items, α=0.83) and perceived good health (MOS-3, 3 items, α=0.89);

–Nottingham Health Profile (Bucquet et al., 1990, Hunt and Wilkund, 1987) measured functional limitations (NHP-1, 13 items, α=0.92) and psychological distress (NHP-2, 14 items, α=0.88).

4. Results 

4.1. Descriptive data 

Global data from medical reports and psychological questionnaires are given on Table 3 for the two groups of patients (IP and CLBPP). Only few differences were found between the groups: scores significantly differed in causal attribution of pain onset, irrational beliefs, distraction, perceived good health, and of course number of GP visits pain, days on sick leave and functional limitations. Anxiety and depression levels as assessed by the STAI-state and the CES-D scores were similar in both groups at inclusion time T1, and were slightly lower at follow-up time T2 also in both groups, without any significant difference.

Table 3. Distribution of the main data in sample

		Improvement		Chronic pain
		T1	T2	T1	T2
	N	60		30
	Age	41.9(11.6)		42.8(11.2)
	Predispositional factors
	STAI-Y trait	40.2 (10.5)		37.7 (10.3)
	CESD - trait	30.2 (9.2)		27.3 (8.9)
	LCS	19.6(2.8)		18.3(3.4)
	Transactional factors
	STAI-Y state	43.9(13.7)	39.2 (14.9)	43.6(13.3)	40.0(15.1)
	CESD state	33.6 (8.9)	30.8 (10.9)	32.9 (7.8)	29.1 (10.9)
	CLCS-1, perceived control	17.5 (3.6)		17.8 (4.2)
	CLCS-2, irrational beliefs	8.3∗(2.8)		10.2∗ (3.3)
	CLCS-3, causal attribution	12.6∗ (3.8)		9.1∗ (2.8)
	CSQ-1, distraction	12.6∗ (4.9)		15.7∗ (4.7)
	CSQ-2, catastrophizing	12.1 (4.7)		12.8 (4.7)
	CSQ-3, reinterpreting	6.5 (2.9)		7.1 (3.6)
	CSQ-4, ignoring	8.1 (3.1)		8.3 (3.9)
	CSQ-5, praying	5.6 (3.4)		6.1 (3.5)
	SSQ-1 quality of social support	20.3 (2.8)		21.2 (2.9)
	SSQ-2 social support availability	19.0(15.9)		16.6(18.7)
	VAS	8.5 (1.8)	2.3 (2.7)∗	9.1(1.5)	4.9 (2.4)∗
	Evolution (T2)
	Days on sick leave		43.8∗ (84.1)		148.8∗ (134.8)
	GP visits for pain, N		2.2∗ (3.9)		7.8∗ (7.9)
	MOS-1, functional limitations		15.0∗ (5.0)		22.0∗ (4.8)
	MOS-2, well-being		17.3 (4.8)		15.6 (5.0)
	MOS-3, perceived health		10.5∗ (3.4)		6.6∗ (4.4)
	NHP-1, functional limitations		16.6∗ (3.5)		21.3∗ (3.3)
	NHP-2, psychological distress		17.8 (3.6)		19.0 (4.2)

T1=Time of inclusion; T2=one-year follow-up. STAI-Y=State-trait Anxiety Inventory; CES-D=Center for Epidemiologic Studies Depression Scale; LCS=Locus of Control Scale; CLCS=Cancer Locus of Control Scale; CSQ=Coping Strategies Questionnaire; VAS=Visual Analog Scale; MOS=Medical Outcome Study; NHP=Nottingham Health Profile. See text for details. SD in brackets.

∗ p<0.05.

4.4. Influence of medical, psychological and social factors on transactional processes 

Further multiple regression analyses were performed on each of these four transactional processes (Table 4). Independent variables were all the predispositional factors (Fig. 1). DP was mostly used by men, with low income, low job satisfaction, who had no previous LBP experience and reduced their activity because of pain; these men were not usually anxious and presented an external LOC. People who were usually anxious and depressive, with few children and low income, were mostly using HH. CR was used by men with a preserved life comfort during the acute pain. Finally, PC was mostly used by men with a high education, several children and a preserved life comfort.

Table 4. Factors influencing transactional processes

	Variables	F	R2	Adjusted R2	β
	Distraction-Praying	7.565∗	0.436	0.379
	Gender				−0.22∗
	Number of children				0.11
	Job satisfaction				−0.20∗
	Income				−0.18∗
	Excessive weight				−0.13
	Reduced activity				0.20∗
	Previous LBP episode				−0.21∗
	Trait-anxiety				−0.20∗
	Internal locus of control				−0.29∗
	Helplessness–hopelessness	12.727∗	0.534	0.492
	Number of children				−0.17∗
	Job satisfaction				0.13
	Income				−0.24∗
	Excessive weight				0.13
	Reduced activity				0.11
	Poor life comforta				0.09
	Trait-anxiety				0.24∗
	Trait-depression				0.42∗
	Cognitive restructuration	3.968∗	0.146	0.109
	Gender				−0.24∗
	Income				−0.18
	Previous LBP episode				−0.17
	Poor life comfort				−0.28∗
	Perceived control	3.404∗	0.209	0.148
	Gender				−0.33∗
	Education				0.32∗
	Number of children				0.25∗
	History of trauma over one year				−0.14
	Previous LBP episode				0.13
	Poor life comfort				−0.19∗
	Trait-anxiety				0.15

Multiple regression analyses.

a This variable corresponded to a self-report of strong deterioration of life comfort due to lumbar pain. ∗ p<0.05.

View Large Image Download to PowerPoint

Fig. 1. Analysis model for the hierarchical multiple regressions.

4.5. Influence of predispositional factors and transactional processes on adjustment 

The same method was applied to the two dimensions of outcome we observed at T2: functional and emotional non-adjustment (Fig. 1 and Table 5). Gender, activity, and history of trauma over one year were predictive of FNA: men with history of trauma over one year who reduced their activity because of pain were at risk of FNA one year after. ENA was only related to trait-depression. With regard to transactional processes, DP predicted a low functional adjustment, although PC seemed to influence good functional outcome, but data fell below significance. HH was predictive of a low emotional adjustment.

Table 5. Influence of medical, psychological social features and transactional processes on the outcome of LBP

	Variables	F	R2	Adjusted R2	β
	Functional non-adjustment	10.112∗	0.473	0.426
	Gender				−0.27∗
	Number of children				0.15
	Education				−0.18
	Income				−0.16
	History of trauma over one year				0.27∗
	Reduced activity				0.28∗
	Poor life comfort				0.09
	Emotional non-adjustment	7.225∗	0.261	0.225
	Number of children				0.11
	Excessive weight				0.16
	Previous LBP episode				−0.14
	Trait-depression				0.44∗
	Functional non-adjustment	13.115∗	0.236	0.218
	Distraction-praying				0.46∗
	Perceived control				−0.17
	Emotional non-adjustment	14.949∗	0.148	0.138
	Helplessness–hopelessness				0.39∗

Multiple regression analyses.

∗ p<0.05.

5. Discussion 

Like previous prospective studies, this one emphasized the role of psychosocial factors as powerful predictors of a chronic evolution after LBP (Burton et al., 1995, Fritz et al., 2001, Hasenbring et al., 1994, Klenerman et al., 1995, Linton, 2000). Patients with acute LBP were found to use four main transactional processes to deal with acute LBP, and the role of these four processes in chronic evolution and low quality of life was documented.

However some limitations in the sample collection and methods have to be acknowledged. First, despite participants were enrolled consecutively and prospectively, we found an unusually high rate of chronification: 33% of patients with a poor outcome, persistent pain and low quality of life. This was probably related to the high rate of work-related accidents among the patients with chronic evolution group (χ2=8.88, p=0.003) This outlines the well-known importance of work-related accident and claims in social insurance as risk factors of poor outcome, but prevents to generalize our results to the common population of LBP patients. Second, was our classification system fully accurate? For technical reasons, the interview period was large (10–90 days), and many patients, hoping for a spontaneous recovery without treatment, have been waiting some days before consulting their GP. So we were unable to question the patients just after the onset of pain, and we grouped together in the IP group patients who were very quickly released from pain, others who got relief later in some weeks or months, and six who felt better, then suffered a relapse, and finally improved. Despite all these 30 participants fulfilled the criteria of being free of pain at one year-follow-up, perhaps their initial coping strategies were different. This has to be looked at in a future study. The same with the recurrent LBP patients, whom we plan to look at in a forthcoming work: although these patients were shown in the literature to take benefit from functional restoration programs as well as non-recurrent LBP patients, Evans et al. demonstrated that they used to develop specific skills, for instance dealing easier with the worker compensation system than the non-recurrent injury patients (Evans et al., 2001, Garcy et al., 1996). So that they might rely on different coping strategies, perception of pain control and social support. Third, our study suffered also from limitations in the assessment battery. Although it is often difficult to interpret data about quality of life in patients with acute pain at the time of the interview, it would have been very interesting to collect MOS and NHP scores at baseline, as to compare with data at T2 time. But our battery was already heavy and of long duration (over one hour), and we did not want to bother our patients further more. As a consequence we could not control for baseline MOS and NHP status in the acute stage.

With regard to the ways these patients reacted to acute LBP, four strategies were identified. The first one, DP, corresponded to a mainly cognitive attitude, which aimed at regulating the emotional content of pain. Patients who used it tried to shift their attention from pain and to think of other things. Patients with high scores on DP expressed irrational beliefs towards pain. They were thinking that the onset and course of pain were depending on Destiny and/or God’s will. They tried to cope with pain by praying, avoiding thinking of pain, and by willingly trying to focus their attention on other things. This strategy may be seen as an adaptive one in acute pain, but seems to play a dysfunctional role on LBP adjustment (Dozois et al., 1996, Rosenstiel and Keefe, 1983, Spinhoven et al., 1989, Riley et al., 1999). In our study, DP was mainly used by men, with external LOC, low trait-anxiety, low income and low job satisfaction, who reduced greatly their activity during acute pain and already had experienced LBP by the past. Patients reacting to pain with the second factor, helplessness–hopelessness, were fully submerged by negative emotions; they were unable to control their feelings of helplessness. They did not believe in cure and recovery. This emotional reaction seems to be specific of dealing with health problems, and has previously been identified in other pathologies (Nuissier, 1996, Tastet, 2001). According to Schermelleh-Engel et al.’s model of pain regulation (1997), the onset of pain makes people overwhelmed with negative emotions. This is in agreement with many previous works emphasizing the well-known role of anxiety and depression in pain perception (Kessler et al., 1996, Lefebvre, 1981, Polatin et al., 1993). Another negative cognition, catastrophizing, has been found associated with a chronic evolution of LBP (Burton et al., 1995, Linton et al., 1999, Picavet et al., 2002). A strong correlation exists between catastrophizing and depression in chronic pain patients (Boothby et al., 1999, Jensen et al., 1994, Rosenstiel and Keefe, 1983, Sullivan et al., 1995). Therefore catastrophizing might be found included in the factor HH. However in our study Catastrophizing, as CSQ-2, was dropped out by the first PCA. So it was impossible for us to look more in detail at its role. We only note that catastrophizing was significantly related to HH (r=+0.41, p<0.000). In our study the emotional reaction to pain was mainly used by persons with high levels of trait-anxiety or trait-depression, which seemed logical. These two transactional processes, DP and HH, were found to have a rather negative influence on the outcome.

The third strategy, CR, associated ignorance and reinterpretation of pain sensations. Patients who were using CR tried to give a new, positive meaning to their pain, and to go on with their life as if pain was absent. This strategy associated strong cognitive strategies applied to pain. It has been clearly identified, under different labels, by previous works on LBP coping but its influence on LBP is not clear (Dozois et al., 1996, Rosenstiel and Keefe, 1983, Spinhoven et al., 1989). In our study, it failed to predict emotional and functional adjustment at 12-month follow-up: the use of CR during the acute LBP episode did not seem to influence the evolution of LBP one year later.

The last strategy, PC, was used by people who had feelings of fair and highly available social support and tried to have an active influence on their sensations of pain and tried to control them willingly. This seemed to be a positive attitude, as patients were feeling that they might influence evolution of pain, and had a positive perception of the help they might receive from their family, friends or colleagues. Like Pellino and Oberst (1992) and Haythornthwaite et al. (1998), we observed that PC was mostly used by highly educated persons. It seemed to have some positive influence on the outcome as well: to think that one can control pain during acute LBP improved functional adjustment at 12-month follow-up. However this trend did not reach significance.

With regard now to outcome, we found that chronic evolution could be related either to a functional or to an emotional non-adjustment. FNA involved functional items of quality of life, duration of sick leave, and GP’s opinion. ENA logically involved state-anxiety, state-depression, and lack of well-being feelings. This seems very consistent, and allows us to go beyond an over-simple approach of the LBP issue, which would only take into account the subjective feeling of pain, as assessed for instance by visual analog scales.

Medical data, psycho-social features and two transactional processes were found predictive of the two adjustment factors. A great amount of literature has been devoted to demographical and medical predictors of the LBP issue: female gender, excessive weight, job conditions, spinal pathology such as spondylolisthesis or scoliosis have been questioned. Our study provides some news in the field: FNA was predicted by male gender, number of children, low income, low educational level, reduction of activity, poor life comfort, and history of trauma over one year. However male gender, history of trauma over one year, and reduced activity were the unique significant predictors in the present multidimensional model. An interesting point is the link between low activity at T1, during the acute pain, and a chronic evolution of pain: our study confirmed that people who stopped working or moving for a long time did not improve, but on the contrary, experienced more often chronic pain than others. Rest is noxious for the back, as it is now well-known, and it has been at the origin of functional restoration programs (Malmivaara et al., 1995, Mayer et al., 1991).

From a psychological point of view, despite depression and anxiety are well-known to be associated with chronic pain, we did not find higher scores on state-anxiety and depression scales in CLBPP than in IP. However our data fall in agreement with previous works, as they identify trait-depression as the main factor of both functional and emotional low adjustment (Clauw et al., 1999, Epping-Jordan et al., 1998, Pincus et al., 2002, Polatin et al., 1993). But our study allows to go further, as looking at relationships between these factors and transactional processes. With regard to the influence of these processes, helplessness–hopelessness had a direct negative effect on emotional adjustment, and DP had a direct negative effect on functional one. PC seemed to prevent from a dysfunctional evolution of LBP.

For the future, important issues would be to provide a better understanding of how LBP patients cope with their problem at the end of the acute phase, and to elucidate to what extent prevention issues might be derived from our findings. Indeed chronification of LBP is a complex and multifactorial problem, and multidimensional studies provided evidence that persistent pain would be best predicted by a combination of somatic, psychological and social parameters (Hasenbring et al., 1994). From a therapeutic point of view, all programs relying either on physical treatment only, or on psychological approaches like cognitive-behaviour therapy in isolation failed to prevent acute LBP from becoming a chronic disability problem. Clearly, the spine and the mind are to be treated simultaneously (Polatin et al., 1993). Comprehensive functional restoration treatment programs are now available, which help chronic LBP patients to reduce their disability by enhancing their physical capacities (Brady et al., 1994, Curtis et al., 1994, Mayer et al., 1994) while providing simultaneously psychological support targeted to self-acceptance of pain, training to problem-solving, reducing catastrophizing and fear avoidance (Chaory et al., 2004, Jousset et al., 2004, Kholes et al., 1990). Prevention programs derived from functional restoration concepts are also developing. They include usually identification of risk factors and patients with high-risk, then early active intervention before LBP becomes chronic: fear-avoidance-based physical exercises, problem-solving therapy, and psychological considerations of negative beliefs such as catastrophizing and kinesiophobia (Fritz et al., 2001, George et al., 2003, Picavet et al., 2002). Evidence of their efficacy has been provided (Gatchel et al., 2003, Van den Hout et al., 2003). To what extent identifying ineffective coping attitudes, and trying to make patients to use more accurate coping strategies would add a supplemental value to such programs remains controversial. Some multidisciplinary prevention programs were found to change significantly negative cognitions and coping attitudes, especially PC, along with an improved outcome (Burns et al., 2003, Jensen et al., 2001). In others, the coping strategies demonstrated little to poor change, with a poor prognostic relevance (Pfingsten et al., 1997). In their recent works Kole-Snijders et al. (1999) and Spinhoven et al. (2004) remained doubtful: in comparison with a control group with passive treatment, cognitive coping skills treatment improved internal pain control and decreased scores for catastrophizing and external pain control as well as operant-behaviour therapy did, but not better: pain coping was not enhanced following treatment. In our study, only helplessness–hopelessness had a clear effect on the emotional issue of low back pain. None of the other transactional processes we looked at had a straight effect on emotional adjustment.

As to conclude, psychosocial factors and transactional processes do play a role when LBP is likely to become a chronic disability. Emotional distress like helplessness–hopelessness and negative beliefs seem to be the most important factors. Future research are needed to provide a better understanding of how acute LBP becomes a chronic disability, and to determine wether early prevention multidisciplinary programs should include psychological intervention on coping strategies.