Materials and Method
In order to meet inclusion criteria for the study, candidates had to have a self-reported diagnosis of chronic back pain, had to be under treatment for back pain, and had to be available at home for 42 consecutive days and nights. The exclusion criterion was a diagnosis of a sleep disorder such as obstructive sleep apnea, a potential confounding factor. The goal was to recruit 90 new participants in order to reach statistically significant conclusions.
Three investigational centers of different specialties (chronic pain rehabilitation, physical therapy, and alternative medicine) participated in the study. Each center secured the approval of its institutional review board. The investigators recruited patients with chronic pain from their patient panels and through newspaper advertising. The ads indicated that the study was on chronic low back pain and the answering machine message indicated that the study method involved sleep surfaces and that pain and sleep would be measured. Following determination that the candidates met the eligibility requirements, the investigators gave patients information about the study and obtained informed consent. Each participant agreed to fill out the study data forms and was promised a small stipend upon study completion.
A commercially available pillow top airbed was used in this study (Figure 1). This sleep system incorporates a pump so that the air pressure is adjustable; the air chamber and multiple layers of foam are designed to minimize areas of excessive skin pressure. The user adjusts the pressure within the air chamber to maximize comfort. The airbed has a firmness control with a range of 0 to 100. The manufacturer states that the typical innerspring mattress has a firmness comparable to the range of 70 to 100 on the airbed. The mattresses used in this study were not marked with the name of the manufacturer or brand name.
Figure 1. Adjustable airbed design.
The airbed was compared with each participant’s own mattress. We made no attempt to gain specific information about the design or performance (firmness) of the innerspring mattresses. The manufacturers use model names that differ for each distributor (eg, furniture store). The inability to identify the mattress design without the cooperation of the many mattress companies made grouping by mattress design — such as the number or firmness of coils — impossible without destructive analysis of the participant’s mattress.
The investigators had a goal of defining the outcomes of sleeping on an airbed vs a control of sleeping on the patient’s own innerspring mattress. Because there has been no gold-standard mattress for chronic back pain, the study used each patient’s mattress as a control. The study incorporated an A-B-A design. First, the participants contributed baseline data from their recent experience on their own mattress. In the second (B) segment, the airbed was installed for 28 days. After the airbed intervention, the patient’s own bed was reinstalled and the patient outcomes tracked for an additional 14 days. There were therefore 3 major periods of study: before, during, and after sleep on the airbed. A-B-A and crossover study designs have previously been used to evaluate beds and pillows.[5,9,10,11]
The investigators developed a protocol that included condition-specific outcome measures and a general health status survey. The investigators selected pain and sleep VAS to be collected daily. In addition, the patient’s diary included the daily airbed firmness control setting (FCS). The patient was required to complete the SF-36 health status survey and the Epworth Daytime Sleepiness Scale (to screen for obstructive sleep apnea) on days 0, 28, and 42. We conducted a final interview to elicit participant comments about both bed designs and bed preferences.
The measures of musculoskeletal function and pain outcome used in this study have been validated in a number of investigations.[15,16,17] Pain and sleep VAS data were recorded for each day of the study. The VAS scales were anchored (annotated) with a description that zero (0) on the pain scale represents a very low level of pain and 10, the maximum, equals a very high level of pain. The sleep VAS scale annotation said that 0 equaled a very low level of sleep quality and 10 represented a very high level of sleep quality. The number was measured in millimeters from the scale; 160 mm was the maximum value. In order to simplify the analysis, the pretreatment values were compared to the average of nights 3-26 (during airbed use) and to the average of nights 31-40 (after the patient was back on the innerspring mattress).
The SF-36 information is reported in 8 dimensions: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. Each dimension is scored with a number between 100 (perfect score) and 0, the worst possible value. For each dimension, the developers of this outcome measure have established a norm value. The norms have been developed by age, so that an individual or group can be compared with norms for the same age group. Previous studies have identified bodily pain and physical functioning as the most responsive dimensions for measuring improvement in chronic back pain.
The Epworth Daytime Sleepiness Scale is reported in points with a range of 0 (no sleepiness) to 24 (the most severe sleepiness). Less than 10 points is considered normal, 10-13 is considered mild, 14-17 moderate, and 18-24 severe. This scale has previously been shown to be helpful in measuring the sleepiness impact of obstructive sleep apnea (OSA), but has not been identified as a measure of sleep disturbance due to chronic back pain. This test was included to help screen out participants who might suffer from undiagnosed OSA.
The study design included analysis of each comparison using matched-pair two-tail t tests to calculate the associated P value. Confidence intervals were calculated for each major outcome measure. Effect size was used to identify the most responsive measures of outcomes. Effect size is defined as the average change in a measure divided by the standard deviation of the baseline measurement, a signal to noise ratio. Analysis of data correlations (such as the firmness control setting vs pain level) was then performed using the most responsive outcome measures.
The pilot A-B-A study had demonstrated an improvement in pain VAS scores that was significant at the level of P < .0003 (1-sided matched-pair t test). The pilot study included 38 patients. The current study recruited 90 new participants in order to allow stratification by indication and by bed preference with a reasonable likelihood of statistically significant results.