Table 1.
Included studies dealing with the transition from subacute to chronic low back pain
| First author, country, year | |||||
|---|---|---|---|---|---|
| Carey [4], USA, 2000 | Dionne [10], Canada, 1997 Training sample and validation sample |
Dionne [9], Canada, 2005 | Epping [13], Jordan, USA, 1998 Same population as in Williams [47] and Wahlgren [46] |
Hansson [14], Denmark, Germany, Israel, Netherlands, Sweden, United States, 2000 | |
| Na, rate at baseline, rate at follow up | 1639, ?, 76% (at 22 months) | 1213, 72%, 84% Training sample (TS): 569, 83% Validation sample (VS): 644, 85% |
860, 68%, 85% Validation study of the final model built in Dionne 1997 [10] |
138, ?, 55.7% | Denmark: 565,?, 87% at 1 year, 78% (at 2 years) Germany: 410,?, 72%, 63% Israel: 327,?, 88%,81% Netherlands: 427; 92%,87% Sweden: 539; 84%,73% US: 484,?, 85%, 79% |
| Study participants | 208 physicians, chiropractors and mid-level practitioners at a group-model Health Maintenance Organization. No data available for the missing 24%. Data for the 76% at 22 months follow up: Mean age 42.5 (functionally better group at 12 weeks), 44.6 (others), 44% male; 24% sciatica at baseline, 83% employed at baseline, 30% on workers compensation at baseline, 17% first time LBP. |
Health Maintenance Organization setting TS: Mean age 46.7 (SD 14.3), 69.6 % over 12 years education Not all sick-listed (only 3.5% received financial compensation at that time) VS: Mean age 46.5 (SD 14.8), 69.3% over 12 years education, Low socioeconomic status was underrepresented. |
Primary care setting Quebec Mean age 38.7, 58.5% male. Only 1.7% were non-white. 44.3 % “only high school‘’, 72.1% married or living as married, 38% current smokers. |
Systematic inclusion of patients in a closed healthcare system (Naval Medical Centre), 100% men, mean age 31.9 (SD 7.0) | Patients on job with sickness and/or disability benefits. Sick listed 100%,age 18 to 59 years, |
| Independent variables: Instruments used | Pain (0–10), gender, race, income, first episode of LBP, duration of LBP before initial visit, employment status, workers compensation, Roland Morris, sciatica, care received, medication used, satisfaction with care | Job factors (employment status at one month, full or part-time student, physical demands of the job (reaching, handling), responsibility for housekeeping, years employed at current job, D.O.Tb.). Psychosocial factors (coping catastrophizing scorec), self-reported health status, stress from family pressures in the past six months, (SCL-90 Depr./ Somat. d). Injury factors (financial compensation, injury diagnosis). Medical factors (migraine, baseline disability (Roland-Morris score), self-reported type of episode, years since onset of back pain, Grouped chronic pain scoree, nr. of associated diagnoses, nr. of pain complaints, present intensity of pain, worst intensity of pain in the past 6 months (VAS), subject ever kept from full-time work for back pain, self-reported injury diagnosis, chest pain in the past 6 month, number of surgery for back pain). Demographic factors (Gender, age, height, BMI, ethnic, marital status, education, income). | SCL-Depression, SCL-Somatizationd | Descriptor Differential Score (DDSg), Sickness Impact Profile (SIPh), Beck Depression Inventoryi | Hannover ADLj, Von Korff pain intensity scalek, SF.36: Vitality, mental health, social functioning and general healthl. Karasek Theorell’s demand-support-control scalem, age, gender, native language, treatment received before sick report, psychological demand, decision latitude (control), physical job demands, surgery during the first year, manipulation/traction, heat and cold (treatment), massage, TENS, physical therapy, back-school program, acupuncture, |
| Dependent variables: Instrument used | Modified Roland Disability Scale at 3 months and 22 months. | Modified Roland-Morris at 2-years. | Functional status (Roland-Morris Disability Questionnaire) at 2 years | 6 and 12 months pain intensity, 6 and 12 months disability, 6 and 12 month depressive symptoms, | Work status at 1 year and 2 years Disability (Hannover ADL) at 1 year and 2 years |
| Definition of non-specific LBP | LBP, no previous care for the current episode, no previous spine surgery, no non-skin malignancy, not pregnant at the time, and able to speak English. Patients were assessed within 2 weeks after medical visit, and after four weeks. | Eligible subjects were adult patients aged 18–75 years who consulted on of the HMOs primary care physicians for a back pain episode. Subjects enrolled at GHCf for less than one year and those who visited walk in care facilities and emergency rooms were excluded. | Eligible if back pain was non-specific and patient had caused at least 1 day of absence from job. Excluded were: pregnancy, serious comorbidity, only cervical pain, or specific cause. | First onset of back pain (thoracic 6 or below), present on a daily basis for 8 (+-2) weeks. | LBP, excluded were: back surgery during the preceding 12 months, back problems caused by fractures, inflammation, infection or tumours |
| Risk assessment time-point | Only the predictors assessed at 4 weeks were used for this review. | 4–6 weeks after the index visit | 3 weeks after their index medical consultation | 6–10 weeks after onset of first time LBP | After 90 days on sick-leave |
| Outcome statistic variables selection process | Logistic regression | Multiple regression for continuous outcome, stepwise, minimum significance level set at 0.01 Recursive partitioning technique for the dichotomized (50%) modified Roland-Morris |
Recursive partitioning with the model developed in Dionne 1997 | Hierarchical Multiple Regression Analysis | Multiple linear regression (Disability) |
| Karjalainen [20], Finland, 2003 | Klenerman [23], UK, 1995 | Leroux [25], Canada, 2004 RAMS-prognosis study |
Loisel [27], Canada, 2002 | Truchon [36], Canada, 2005 | |
|---|---|---|---|---|---|
| Na, rate at baseline, rate at follow up | 134, 81%, 97% to 98% | 162, ?, 100% | 1007, 68%, 84% | 104, ?, 87% | 439, 49%, 73% |
| Study participants | From 36 primary health care centres, 350 general practitioners. Mean age 44 (SD 8.8, range 25 to 60), 58% women, 22% blue-collar workers, satisfaction with work (0 to 10): 7.3 (SD 2.3) 34% with high school diploma, 69% with radicular symptoms below the knee |
Patients from general practitioners (gender for this subgroup not available) |
Workers seeking medical consultation in primary care setting. Baseline characteristics only for patients for which follow up was available: 483 men (mean age 39.4; SD 10.3); 366 women (mean age 38.4; SD 10.9) |
Recruited from all workplaces with more than 175 workers and located within of 30 km from the back clinic. Randomized study, 53,9% men, mean age 40.2 (SD 9.1). Mean days of absence from work 40.6 (SD 14.1) | Workers on sick leave by the Quebec Workers’ Compensation Board Mean age 39 years (range 18 to 60) 56% men, 63 % married or part of a couple, 60% had high school diploma, 25% higher diploma |
| Independent variables: Instruments used | Gender, age, BMI, high school diploma, blue-collar worker, duration of sick leave at baseline, radicular symptoms below the knee, intensity of pain at baseline (0–10), Oswestry Disability Index, perceived risk of not recovering (0–10), expectation regarding the effectiveness of treatment (0–10), satisfaction with work (0–10), self-rated health status for age | Age, sex, referring doctor, marital status, employment status, smoking habits. Previous and present history and severity of LBP, ratings of the four fear-avoidance contextual variables: stressful life events with rating scale Holmes and Rahe, Somatic symptoms (MSPQn), Previous Pain Historyo, Pain Coping Strategies (Slade)p. Pain Drawing, Modified Zung, Disability (Oswestry) Non-organic symptoms, non-organic signs, body mass index, neurological tests, SLRq, Prone knee bend, Hip flexion, lateral flexion, sagittal movement, area affected, clinical diagnosis by doctor. |
Age, marital status, education, work schedule, duration of work in same job, stressful life events in past 12 months, baseline Roland-Morris score, self-reported type of episode, psychological demands and job decision latitude (Karasek’s Job Content Questionnairer), somatization score, depression score, self-reported health status, fear-avoidance beliefs about back pain, likelihood of loosing job within 2 years, psychological job demands, worst intensity of back pain in past 6 month, physical workload (level of physical effort x lifting heavy load, whole-body vibration, frequency of trunk flexion, lateral bending) Combinations of psychological demands and social supports at work (work APGARs), job satisfaction |
Quebec Task Force Classification (QTFCt). | Coping (CPCIu), Disability (Roland-Morris disability questionnaire), Pain (NRS-101v), Depressive Mood (Hospital Anxiety and Depression Scale HADS), Catastrophizing (Catastrophizing scale of Coping Strategies Questionnaire) |
| Dependent variables: Instrument used | High Oswestry Disability Index (dichotomized) Intensity of Pain (0 to 10), daily symptoms, bothersome Pain, Pain interfering with work or daily Life, HRQL (15D), satisfaction with medical care (0 to 10), health care cost, sick leave (0, 1 to 30, or over 30 days). At 12 months. |
Level of pain, Disability, whether off work, course of LBP at 12 month For this review, only the analyses for disability are considered (because only these predict from the 2 months assessment) |
Roland-Morris score at 1 year | Sickness Impact Profile (SIPh), Oswestry Questionnaire, McGill Pain Questionnaire (MPQ) Work status, assessed from patient questionnaire (return to the identical work performed before onset of back pain). |
Pain, Disability, Depressive mood at 6 months, work status |
| Definition of non-specific LBP | LBP (with or without sciatica), excluding: need for operative treatment, pregnancy, history of specific back disease (cancer, fracture, spondylarthritis, or infection), somatic or psychiatric disorder preventing rehabilitation, substance abuse, consultation with a specialist in physical and rehabilitation medicine during the past year, inpatient rehabilitation for back pain during the last 3 years, 3 month of continuous sick leave during the preceding year, and impossibility of a work visit. | Benign, musculoskeletal LBP. | Non-specific back pain (including the thoracic, lumbar, and lumbosacral areas), at least 1 day of sick leave. Exclusion: specific pathologies, such as cancer, spinal infection, vertebral fractures, systematic diseases, cauda equine syndrome, referred pain, pregnancy, and major medical illness that could affect work status. | First four categories of the Quebec Task Force Classification (QTFCt). Pregnant, spine fracture, significant degenerative spine disease (spondylolisthesis, Grade 2 or more), non-mechanical spine disease (tumour or infection), or major comorbid condition were excluded. | On sick-leave between 3 and 12 weeks after the accident. Exclusion criteria: insufficient understanding of French, return to work, pregnancy, previous back surgery, severe spinal pathology like fracture, tumour, infection, cauda equine syndrome, and symptoms suggesting nerve compression. |
| Risk assessment time-point | After 4–12 weeks of LBP making working difficult | For this review, we included only analysis from the 2 months assessment | About 3 weeks after index medical visit | 4–12.5 weeks after initial medical visit | Between 3 and 12 weeks on sick leave |
| Outcome statistic variables selection process | Generalized estimating equations (GEE) (repeated measures data, with mean response modelled as a logistic regression model). | Principal component analyses Multiple linear regression |
Analyses of Covariance with adjustment of the baseline Roland-Morris score and adjustment for confounding factors with multiple regression. | Logistic regression | Hierarchical multiple regression |
| Van der Weide [39], Netherlands, 1999 | Von Korff [42], USA, 1993 | Von Korff [43], USA, 2005 same population as Von Korff 1993 und Dionne 1997 |
Wahlgren [46], USA, 1997 Same population as in Epping-Jordan 1998 and Williams 1998 |
Williams [47], USA, 1998 Same population as in Epping-Jordan 1998 and Wahlgren 1997 |
|
|---|---|---|---|---|---|
| Na, rate at baseline, rate at follow up | 120, 85%, 90% | 1213, 72%, 94.4% | 1213, 72%, 93% (1 year), 84% (2 year)%, 68% (5 year) | 138, ?%, .55% | 138, %, 60% |
| Study participants | Patients on sick leave, health care and university workers. 33% male, mean age 39 (SD 8.7). Smokers 33%, sporting activities more than one hour per week 68%, suspicion of nerve root compression 24%, radiating pain until knee 24%, beyond knee 38%, on sick leave during last year 31%, on sick leave more than once 44%, mentally demanding work 25%, mixed mentally or physically 53%, physically demanding work 22% |
Health Maintenance Organization (HMO), age 18 to 74, 47.1% male, 52.9 female, 33.2% college graduate, 61.6% high school graduate, 5.2 less than 12 years of education | Health Maintenance Organization (HMO), age 18 to 74, 47.1% male, 52.9 female, 33.2% college graduate, 61.6% high school graduate, 5.2 less than 12 years of education | Naval Medical Center, San Diego, 100% men, mean age 30.5 (SD7.3), mead education 12.8 (SD 1.8) years, 69% married 52% pain without radiation, 18% pain with proximal radiation, 13% with distal radiation, 13% pain with radiation and neurological signs, 2 patients with compression of spinal nerve root confirmed by specific imaging techniques. |
as above |
| Independent variables: Instruments used | Demographic factors, perceptions of work conditions, low-back pain characteristics, pain intensity (VASw), functional disability (Roland Disability Questionnaire), general health perception (Nottingham Health Profile), coping, health locus of control | Demographic, pain (characteristic pain intensity=average of three 0 to 10 ratings: back pain at the time of the interview, average pain, and worst pain during previous 6 months), Disability score, Disability days, Days in pain, recentness of onset (time since the first episode of back pain), Pain grade (five categories). SCL-90-R depression and vegetative symptoms scale. | Demographic, pain (characteristic pain intensity=average of three 0 to 10 ratings, back pain at the time of the interview, average pain, and worst pain during previous 6 months), Disability score, Disability days, Days in pain, recentness of onset (time since the first episode of back pain), Pain grade (five categories). SCL-90-R ddepression and vegetative symptoms scale, number of pain sites | Descriptor Differential Scale (DDSg), Sickness Impact Profile (SIPh), Beck Depression Inventory (BDIi) | Job satisfaction (Job Descriptive Index), modified work APGARs, Pain VAS, Descriptor Differential Scaleg, Sickness Impact Profileh, Quality of Well Being Index, Hamilton Rating Scale for Depression, Beck Depression Inventoryi, Automatic Thoughts Questionnairew, Orthopedic impairmentx. |
| Dependent variables: Instrument used | Roland-Morris Disability Questionnaire at 3 and 12 months. Return to work status at 3 month, 6 month and 12 month, Sick leave at 1 year, Time to return to work. |
Poor outcome (Grade III: high disability-moderately limiting back pain, Grade IV: high disability-severely limiting back pain). At 1 year | Poor outcome (Grade III: high disability-moderately limiting back pain, Grade IV: high disability-severely limiting back pain). At 1 year and at 5 years | Pain at 6 months Disability at 6 months Distress at 6 months Overall Clinical Outcome at 6 months. |
Pain at 6 months Disability at 6 months Distress at 6 months Overall Clinical Outcome at 6 months. |
| Definition of non-specific LBP | LBP for at least 10 days, pain located below the scapulae and above the gluteal fold. Not pregnant. | Patients at a primary care back pain visit, visits to emergency room and walk-in care facilities were not included. | Patients at a primary care back pain visit, visits to emergency room and walk-in care facilities were not included. | First onset back pain (T6 or below) present on a daily basis for the previous 8 (±) weeks. Excluded were prior episodes of back or other pain on a daily basis lasting one week or longer; taking medications known to affect mood (e.g. antidepressants, anxiolytics), major surgery within the preceding 12 months, and back pain secondary to neoplastic disease, osteomyelitis, of fracture. | First-onset LBP Exclude were: prior episode of back or other pain on a daily basis lasting for 2 weeks or longer, major medical illness (e. g, insulin-dependent diabetes, chronic obstructive pulmonary disease, taking medications known to affect mood (e. g, antidepressants), prior back surgery, pain secondary to neoplastic disease, osteomyelitis, or fracture). |
| Risk assessment time-point | After 10 days on sick-leave | 3–6 weeks after index medical visit | 3–6 weeks after index medical visit | 6–10 weeks after onset of first time LBP | 6–10 weeks after onset of first time LBP |
| Outcome statistic variables selection process | Logistic regression All variables with P < 0.15 in the univariable analysis were included in the multivariable analysis. Also included were variables with practical relevancy for the occupational physician. Interaction terms were evaluated |
Logistic regression | Latent Transition Regression Analysis (LTRA) Risk score estimation |
Change from baseline to six months: cross-tabulation | Principal Components Analysis, Hierarchical multiple regression |
a Sample size. ? means not reported, LBP low back pain.b D.O.T.: Dictionary of Occupational Titles of the US Department of Labor.c Coping Catastrophizing Score: they used a score based on selected items from the Vanderbilt Pain Management Inventory and the Coping Strategy Questionnaire.d SCL-90 (Depression / Somatization): Dionne et al. (1997[10]) used selected items from the SCL-90 Depression and Somatization subscales.e Grouped Chronic Pain scores: pain grade based on characteristic pain intensity, disability score and number of pain disability days in the past six months).f GHC: Group Health Cooperative of Puget Sound, Health Maintenance Organization (HMO). g DDS: The Descriptor Differential Scale is a self-report measure of current pain intensity and unpleasantness.h SIP: The Sickness Impact Profile is a 136 self-report questionnaire that evaluates the degree to which pain interferes with usual daily activities. i BDI: Beck Depression Inventory, 21-Item self-administered questionnaire that assesses severity of depressive symptoms. j Hannover ADL: evaluates function limitations due to LBP (12 items), score from 0 (greatest possible limitations) to 100 (no limitations). k Von Korff pain intensity scale: score between 0 (no pain) to 10:l Vitality, mental health, social functioning and general health, subscales from SF-36, a health-related quality-of-life instrument. The scores for all the subscales range between 0 and 100 (higher values = better function).m Karasek Theorell’s demand-support-control scale: A questionnaire evaluating “job strain”, psychological demands, conflicts at work, and the possibility to utilizing an individuals abilities and the possibility to develop new skills.n MSPQ: Modified Somatic Perception Questionnaire, 13 Items for somatic symptoms.o Previous Pain History: scale for previous pain, 1 for no pain at all, 10 for worst imaginable pain, three categories of pain: internally produced pain (headaches, toothache, etc.), externally produced pain (fractures, dentistry, etc.) and accidental pain (cut finger etc.). p Pain Coping Strategies: five choices of strategies (Medication, rest, doctor, physical exercise, ignoring).q SLR: Straight Leg Raise.r Karasek’s Job Content Questionnaire: 9 Items for psychological job demands, 9 Items for job decision latitude. s Work APGAR Scale: adaptation, partnership, growth, affection and resolve. 17 Items to evaluate the perception of support at the workplace and job enjoyment (in the Leroux study the score was dichotomized by the median). t Quebec Task Force Classification (QTFC): a commonly used classification based on simple clinical criteria including signs and symptoms (pain and neurological examination data), imaging test results, and response to treatment (in this study only QTFC categories 1–4 were used; 1 = Pain without radiation, 2 = Pain with proximal radiation (above knee), 3 = Pain with distal radiation (below knee), 4 = Pain with distal radiation and neurological signs. u CPCI: self-report measure, rates frequency of use of behavioural and cognitive coping strategies (guarding, resting, asking for assistance, relaxation, task persistence, exercise/stretching, coping self-statements, and seeking social support). Four answer categories-version used. v NRS-101: Numeric Rating Scale 0 to 100. w VAS: Visual Analogue Scale. wAutomatic Thoughts Questionnaire: 30 statements, score is reflective of distressing cognitions and are related to sub-clinical and clinical depression. x Orthopaedic impairment: Waddel Impairment Index for measuring the severity of orthopaedic disease in back disorders. Score from 0 to 50 (more impairment)
Tr: “training sample” (to define the model), Va: “validation sample”: They randomly divided the population in two samples: one to develop the model, the second to validate the model