Neck and back pain are common disorders around the globe. According to the National Health Information Survey of 2002 in the United States, over a 3-month period prior to a survey, 26% of surveyed individuals experienced lower back pain and 14% experienced neck pain (Lethbridge-Çejku et al., 2004). Some 5–6% of patients with back pain develop chronic disability. These patients consume 80% of health care resources spent on treating back pain (Hashemi et al., 1998). The need to develop effective treatment options for patients with chronic pain conditions is a major challenge for health care professionals involved in musculoskeletal/pain management. Good and accurate outcome measures are critical for evaluation of treatments for neck and back pain.

Chronic neck and back pain results in multidimensional dysfunction involving physical function, affective and mood state, activities of daily living, patients’ satisfaction with life in general, quality and quantity of pain, and sleep and fatigue. These result in significant decrement in a patient’s quality of life (QOL). The outcome measure should be able to evaluate all these aspects of a patient’s functioning. In addition, the measure should be able to track these over time. Tracking allows comparison of how treatment or life events impact a patient’s function and quality of life.

The main purpose of the LIFEwar^SM system (Granger et al., 1995; Baker et al., 1996) is to measure QOL in patients from teenage through adult life. The domains measured in the example that follows are physical function, mood and affective state, and pain. The measures were developed using Rasch analysis (Wright and Stone, 2004). To understand the LIFEware measures, it is important to review key concepts of Rasch analysis.

Rasch Analysis

Usually items are rated with numbers to indicate more or less of the trait that is presumed to be homogeneous. Rasch analysis permits the rating of a limited set of attributes that are representative of the underlying trait. Whether observed or self-reported, the sum rating of the attributes represents how much of the trait has been mastered.

The purposes of Rasch analysis are to maximize the homogeneity of the trait and to allow greater reduction of redundancy with no sacrifice of information by decreasing items and/or rating levels to yield a more valid and simple measure.

To develop a scale to measure the performance of a particular task, we need to break down the task into its principal components. Successful completion of a task depends on two factors: the person’s ability and the difficulty of the task. If the person’s ability exceeds the difficulty of the task, then the person will successfully accomplish the activity. In order for an observation to have contextual meaning, the concepts of a person’s ability and the difficulty of the task must each have independent directions; otherwise, an observation is devoid of meaning. It is within the context of direction that ability and difficulty become relevant.

Having independent directions of the two factors permits quantification and comparison with other independent observations. Criteria may be established to determine whether there is sameness or difference; and if there is difference, in what direction along a latent variable or how much distance along that direction.

Only after replication of observations of people with different abilities doing tasks of different difficulties can the variables of ability and difficulty be used to calibrate a yardstick related to performance of a particular task. Both can be measured along the same hierarchy. This is known as conjoint additivity.

The only tool for achieving conjoint additivity specifications—expressed as B_n − D_i (ability minus difficulty)—is the logarithm of the probability (log odds). Rasch measurement is based on probability rather than certainty. In other words, a person-rating can be achieved in a number of different ways. However, depending on the item of hierarchy, only one-way is expected. Rasch analysis measures the degree of expectedness.

Rasch analysis proposes a model for measurement. Data collected are tested against that model to determine whether the data fit the model. If the data are judged to fit the model then the data form a measure.

In summary, Rasch analysis is a mathematical model based on a latent trait and accomplishes stochastic (probabilistic) conjoint additivity (measurement of the item difficulty and the person ability on the same metric).

The Rasch modeling provides a philosophical and mathematical foundation for the theory of objective measurement. It is operationalized by the WINSTEPS software (Linacre and Wright, 2000).

The LIFEware measures for musculoskeletal conditions were tested as the Medical Rehabilitation Follow Along (MRFA) (Granger et al., 1995; Baker et al., 1996). The reliability of MRFA was established in a study of 47 patients. The patients completed the musculoskeletal form of the MRFA instrument on two occasions separated by an interval of 1–7 days. Responses were examined using the intra-class correlation coefficient (ICC) and kappa. ICC values for the sections of the MRFA instrument examining quality of daily living and physical functioning ranged from 0.74 to 0.97.ICC values for items assessing pain and feeling of well-being were more variable, ranging from 0.36 to 0.93. The kappa values displayed a similar pattern. The reliability of the MRFA instrument was found to be adequate for gathering screening information in outpatient settings. The validity of the measure was confirmed by comparing it to the SF-36 (Ware, 1993), comparing pre- and post-treatment ratings and comparing the scale ratings with therapists’ ratings of improvement. The scale includes within it elements of the Functional Assessment Screening Questionnaire (Granger and Wright, 1993), Oswestry Scale (Fairbank et al., 1980), short form McGill Pain Questionnaire (Melzack, 1987), and Brief Symptom Inventory (Derogatis and Melisarotos, 1983).