Prediction of Sickness Absence in Patients with Chronic Low Back Pain: A Systematic Review

Objectives: To provide evidence of predictors for sickness absence in patients with non-specific chronic low back pain (CLBP), distinguishing predictors aimed at the decision to report sick (absence threshold) and decision to return to work (return to work threshold). Methods: Medical and psychological databases were searched, as well as citations from relevant reviews. In- and exclusion criteria were applied. Two reviewers assessed the methodological quality of the papers independently. Results: Many different predictors were studied, and few factors were studied more than once. Consistent evidence was found for own expectations of recovery only as predictor for the decision to return to work. Patients with higher expectations had less sickness absence at the moment of follow-up measurement. As expected, different predictors were found aiming at the absence threshold or the return to work threshold. Furthermore, predictors varied also with the measurement instruments used, timing of follow-up measurements, and definition of outcomes. Until now, too few studies are available to overcome several potential sources of heterogeneity. Conclusions: No core set of predictors exists for sickness absence in general. The characteristics of the study including the decision to report sick or to return to work determined the influence of several predictors on sickness absence in patients with CLBP. Further research and use of a core set of measurements and uniform definitions are needed to predict sickness absence and return to work in patients with CLBP.

Appendices

Appendix 1: Literature search

Each database was searched for: “Low Back Pain” entered as MesH term and free text word (Psychinfo only LBP in title or abstract), combined with MesH terms and free text words. In each database, the same free text words were used. Mesh Terms differed per database. In Table A1 the MesH terms are presented. Some MesH terms were only used in combination with the free text words “work” or “working” or “occupation^*” or “job^*” or “employment” or “unemployment”, presented in the table as combination terms.

Free text words used for the searches:

Return* to work	Off work	Work disability
Job resumption	Absent from work	Work ability
Job loss	Back to work	Work incapacity
Work loss	Work status	Work capacity
Work resumption	Employment status	Occupational disability
Work absence	Occupational status	Occupational ability
Absenteeism	Vocational status	Occupational incapacity
Sick* leave*	Job status	Occupational capacity

Table A1 MesH terms used in the different databases

Appendix 2: In- and exclusion criteria

A paper was included when all inclusion criteria were met and none of the exclusion criteria were applicable; a paper was included for further judgment when the exclusion criteria did not apply, but (some of) the inclusion criteria were not clearly specified. A paper was excluded when one of the exclusion criteria were met.

Inclusion criteria

I1 Study population
1	Non-specific low back pain (LBP) or musculoskeletal pain in which a subgroup of patients with LBP is presented separately. Non-specific = pain (with or without radiation) without specific origin. Low back = the lumbar spine
2	Sub acute, chronic or recurrent LBP
	Sub acute – chronic = complaints and/or sick leave associated with back pain ≥ 4 weeks
	Recurrent LBP = defined as ‘recurrences or episodes of back pain in the previous year’
I2 Design
3	Observational study, (prospective cohort study, follow-up study, longitudinal study) or Randomized Control Trial (with therapy as prognostic factor)
4	Prognostic factors should be identified (including age, gender)
I3 Outcome
5	In a cohort of patients on sick leave at baseline:
	– RTW during study yes/no
	– Duration of sick leave (in days)
	In a cohort of patients working at baseline:
	– Sick leave during study yes/no
	– Duration of sick leave (in days)
	– Number of sick leave registrations

In a mixed population (on sick leave and working) both outcomes can be applicable. These outcomes should be described separately.

Exclusion criteria

E1 Study population
a	Only a group of patients with musculoskeletal complaints other than LBP
b	LBP with a distinct causal diagnose (Bechterew, Rheumatoid Arthritis, spondylolisthesis, fracture, infection, inflammatory process, neoplasm) or a mixed population without making the distinction between non-specific and specific LBP. Degeneration of the spine without nerve compression (spondylosis) is not a specific cause for LBP
c	Patients with cardiovascular or pulmonal disease, hypertension, drug addiction, psychopathology or pregnancy, or a mixed population in which no distinction is made between these groups and the non-specific LBP group
d	Patients in which >10% has a post surgery status in the past 2 years, or a mixed population in which no distinction is made between post surgery patients and non surgery LBP patients
e	Duration of LBP or sick leave ≤ 4 weeks, or a mixed population in which acute, sub acute and chronic patients were not described separately
f	Duration of complaints = ‘New workers compensation claim’
E2 Design
g	Cross-sectional studies without follow-up period,
	retrospective studies,
	(systematic) reviews,
	updates of (systematic) reviews,
	abstracts of congress papers,
	commentary on other papers/letters to the editor
h	Study population < 40 LBP patients (case studies, case reports)
i	No prognostic factors analyzed (including age, gender)
E3 Outcome
j	Outcome other than described above

Appendix 3: Methodological criteria list for cohort studies

A. Was the duration of complaints ≥ 6 weeks?		Yes/No/?
B. Was the description of in- and exclusion criteria adequate? (1 or 2, and 3)	1 Inclusion 2 Exclusion	Yes/No/?
	3 Pregnancy
C. Was the study population/potential prognostic factors described adequately? (all 5)	Age Gender	Yes/No/?
	Duration complaints
	Sick leave
	Surgery
D. Are other potential prognostic factors described adequately? (at least 1)	1 Physical	Yes/No/?
	2 Psychosocial
	3 Work related
	4 Pre-existing
	5 Financial
	6 Socio-demographic
	7 Treatment
E. Was the follow-up duration ≥ 12 months?		Yes/No/?
F. Was the number of dropouts described and acceptable?		Yes/No/?
G. Are demographic/ clinical data described adequately of the dropouts? (at least 1)	Description Statistical test	Yes/No/?
H. Was the outcome defined adequately?		Yes/No/?
I. Are the performed analyses adequately (described)? (1 or 2 or 3 and 4)	1 e.g. T-test	Yes/No/?
	2 e.g. Risk ratio
	3 e.g. Regression analysis
	4 Correction for multiple comparisons

Appendix 4: Methodological criteria list for assessing RCTs (18)

A	Was the method of randomization adequate?	Yes/No/?
B	Was the treatment allocation concealed?	Yes/No/?
C	Were the groups similar at baseline regarding the most important prognostic indicators?	Yes/No/?
D	Was the patient blinded to the intervention?	Yes/No/?
E	Was the care provider blinded to the intervention?	Yes/No/?
F	Was the outcome assessor blinded to the intervention?	Yes/No/?
G	Were co-interventions avoided or similar?	Yes/No/?
H	Was the compliance acceptable in all groups?	Yes/No/?
I	Was the dropout rate described and acceptable?	Yes/No/?
J	Was the timing of the outcome assessment in all groups similar?	Yes/No/?
K	Did the analysis include an intention-to-treat analysis?	Yes/No/?