What is the prognosis of back pain?

Understanding prognosis is important in managing low back pain. In this article, we discuss the available evidence on low back pain prognosis and describe how prognostic evidence can be used to inform clinical decision making. We describe three main types of related prognosis questions: ‘What is the most likely course?’ (Course studies); ‘What factors are associated with, or determine, outcome?’ (Prognostic factor or explanatory studies); and ‘Can we identify risk groups who are likely to have different outcomes?’ (Risk group or outcome prediction studies).

Most low back pain episodes are mild and rarely disabling, with only a small proportion of individuals seeking care. Among those presenting for care, there is variability in outcome according to patient characteristics. Most new episodes recover within a few weeks. However, recurrences are common and individuals with chronic, long-standing low back pain tend to show a more persistent course. Studies of mixed primary care populations indicate 60–80% of health-care consulters will continue to have pain after a year. Important low back pain prognostic factors are related to the back pain episode, the individual and psychological characteristics, as well as the work and social environment. Although numerous studies have developed prediction models in the field, most models/tools explain less than 50% of outcome variability and few have been tested in independent samples. We discuss limitations and future directions for research in the area of low back pain prognosis.

Importance of low back-pain prognosis

Prognosis is a description of the probable course or prediction of the outcome of a health condition over time. Important to prognosis is consideration and assessment of characteristics or factors that are associated with or determine the course of a condition. Health-care professionals may use prognostic information to educate or inform the management of their patients .

Several characteristics of low back pain make consideration of prognosis extremely important. Low back pain is common and costly, with most of the economic and social costs attributed to individuals who have prolonged disability. It has not been possible to identify a specific cause for most cases of low back pain, and interventions with strong evidence of effectiveness have not been identified . Therefore, there has long been attention on prognosis research in the field.

There are three main types of related prognosis questions: ‘What is the most likely course?’, ‘What factors are associated with, or determine, outcome?’ and ‘Can we identify risk groups who are likely to have different outcomes?’ ( Fig. 1 ). Understanding the likely course of low back pain provides a descriptive picture and is important for patient counselling, planning, management and monitoring. Identifying prognostic factors that are associated with worse or better disease outcome (explanatory studies) can help us to understand possible determinants and causal pathways for low back pain, which may lead to effective interventions. We can use prediction models or risk scores to identify groups of patients who are likely to have worse (or better) outcomes; this can assist in clinical decision making and to inform patients more specifically about their likely outcome (outcome prediction studies).

In this article, we discuss the state of the art on prognosis of low back pain, including course, important prognostic factors and identifying high- or low-risk groups. We discuss the current evidence in each of these areas of prognosis, how evidence is used in practice and some future directions in the field.

Evidence on low back pain prognosis

What we know about prognosis of low back pain comes from a large number of studies. Table 1 describes the three basic types of prognosis studies. Essentially, three components define prognosis studies: (1) a clearly defined cohort or population at a common and preferably early point in the course of their condition (e.g., a patient with a new episode of low back pain presenting for care), (2) definition of a future outcome (e.g., recovery from illness) and (3) measurement of the likelihood of this outcome and/or the strength of its determinants. With respect to probable outcomes and important determinants, different ‘phases of investigation’ in prognosis studies provide different strength of evidence: exploration, confirmation (validation) and impact (implementation) studies.

Table 1

Summary of the three different prognosis study types, including brief description of the purpose, the key design features, presentation of results, and interpretation of findings from each.

Type	Ultimate Purpose	Key design features	Presentation and interpretation of study results
Studies of prognostic course	Describe likely outcome	– Clear description of population (or subgroup) at common, early point – Rationale for follow-up period(s) – Descriptive analysis	– Cohort description provides information about generalizability – Studies recruiting people seeking care usually describe clinical course (rather than natural history) – Longer and multiple follow-up periods may give better picture of true course
Prognostic factor studies (Explanatory)	Investigate causal determinants	– Conceptual framework guides study (with sufficient evidence from previous studies, patho-physiological plausibility, and relevance to our understanding or treatment of the disease) – Prognostic factors (including prognostic factors of interest and potential confounding factors) considered are based on hypotheses – Analyses are guided by hypotheses	– Aim to understand which prognostic factors are associated with and/or contribute to low back pain outcome – Presentation of strength and independence of association between factors and outcome – Confirmatory studies test independent association to identify opportunities for possible intervention studies
Outcome prediction studies	Identify risk groups	– Prognostic factors used in model development can come from anywhere (causal determinants are good candidates, but no need to include them; confounding not important) – Analysis aims to identify combination of factors that best predicts outcome – Internal and external validation important – Later phase studies should test impact on practice	– Outcome prediction models or tools are used to identify risk groups – Presentation of calibration and accuracy of outcome prediction model or tool – Ultimately how successfully the model or tool is used in practice to improve clinical management and outcomes