Non-cancer pain (NCP) is a wide-spread debilitating condition with an increasing incidence rate [1
]. The Institute of Medicine (IOM) in 2011 estimated that approximately 100 million Americans were affected by pain, and the cost of its treatment is nearly $600 billion annually [4
]. Besides the rising incidence of NCP, the burden to caregivers parallels the disease burden for the patient with NCP. One study found that 70% of patients with NCP reported higher self-perceived burden, achieving a minimally important clinical difference [5
]. Caregivers of those with NCP also experience significant burden with impairments in mobility, activities of daily living, and other self-care activities as well as overall level of pain being the largest contributors to higher subjective caregiver burden [6
Since improvements in NCP cannot be measured by laboratory values, clinicians have to rely on subjective assessments of pain to derive effectiveness of any intervention designed to reduce pain. Several pain scales have been validated in the measurement of pain [8
]; however, in addition to improvement in pain intensity, patients as well as clinicians hope to see improvements in health-related quality of life (HRQoL). Currently, there is limited long-term evidence on the changes in HRQoL that interventions for NCP may bring. With the current rise in opioid use and abuse [12
], having validated measures of HRQoL could help ensure that an appropriate benefit to risk ratio is being preserved. One study reported that anxiety, depression, and severe impairment were prevalent within this population [14
], and it has been shown that patients with pain of high intensity as well as persistent pain have lower HRQoL [15
]. Additionally, several studies evaluated the impact of an intervention on HRQoL in those with pain [17
To date, only four HRQoL instruments have been validated in patients with pain. Vartiainen et al. found the 15D instrument could discriminate pain-related factors and was sensitive in chronic pain patients [21
]. Luo et al. validated the Medical Outcomes Study Short Form-12, Version 2 (SF-12v2) in chronic back pain alone, but not in chronic pain overall [22
]. The Medical Outcomes Study Short Form-36 (SF-36) and the European Organization for Research and Treatment of Cancer Quality of Life Core Questionnaire (EORTC QLQ-C30) were also validated in patients with chronic nonmalignant pain [23
]. Other measures are available to measure pain severity and functionality. Unidimensional scales such as the Visual Analog Scale or the Numeric Rating Scale measure only pain severity. Multidimensional scales such as the Brief Pain Inventory and the Treatment Outcomes of Pain Survey not only measure pain severity but other dimensions affected by pain such as pain interference with life activities and pain coping style [24
]. Additionally, performances of the EQ-5D and the SF-6D, which are preference-based measures, have been compared in patients with chronic pain. The EQ-5D appears to have better construct validity and responsiveness in this patient population [25
The SF-12v2 is one of the generic instruments that has been widely used to measure HRQoL in patients with chronic conditions [27
]. Since the SF-12v2 is a shorter instrument, it requires less patient and clinician effort [34
]. Additionally, the SF-12v2 is embedded in the Medical Expenditure Panel Survey (MEPS), providing an avenue to evaluate HRQoL among patients with NCP on a population level. There is a need to validate the SF12-v2 to help support its use to quantify HRQoL in NCP patients. Therefore, the objective of this study was to evaluate the reliability and validity of the commonly used SF-12v2 in a NCP cohort derived from MEPS.
Patient-reported outcomes become essential in conditions that do not have objective measures for determining treatment outcomes like NCP. With the prevalence of NCP on the rise [44
], as well as the cost of treating it [45
], HRQoL of patients with NCP can be used to inform clinicians on treatment selections and also to monitor health outcomes. Information on psychometric properties of HRQoL instruments is crucial to encourage and expand the use of HRQoL instruments in real world settings. This study was the first to evaluate validity of the SF-12v2 for use in patients with NCP using a nationally representative US cohort. We found that the SF-12v2 is valid and reliable for quantifying HRQoL for patients with NCP.
Both the MCS12 and PCS12 demonstrated acceptable internal consistency and test-retest reliability. These findings on internal consistency reliability were similar to a previous study in patients with back pain [22
]. To the best of our knowledge, no study evaluating a non-cancer pain population has performed test-retest reliability. Our analysis methods parallel those of Cheak-Zamora et al. who evaluated test-retest reliability of the SF-12v2 in a general cohort of MEPS participants [37
]. Our results of test-retest reliability among NCP cohorts were similar to Cheak-Zamora et al. [37
In evaluating construct validity, the correlations we observed between PCS12/MCS12 and perceived health were similar to Cheak-Zamora et al. [37
]. However, our results in a NCP cohort showed the PCS12 and MCS12 to be more discriminant and convergent with perceived mental health, respectively. Vartiainen et al. showed that pain intensity did not predict the score on the 15D instrument [21
]. However, we found that both the PCS12 and MCS12 in round 2 were able to predict the likelihood of reporting physical and cognitive limitations in round 3, and therefore, should be able to predict future limitations in physical and mental health.
The findings of MCS12 were similar to those found originally by Ware et al. and Gandek et al. [27
]. For concurrent validity, the MCS12, unlike the PCS12, was not significantly lower for each unit increase in the number of chronic conditions but, instead, was only significantly different between those with no and one chronic condition as compared to those with three chronic conditions and those with four or more. The lack of significance between each unit increase in the number of chronic conditions could be due to scale recalibration, where there is either a reprioritization or reconceptualization of quality of life among those with chronic conditions which then effects the MCS12 [46
]. Cheak-Zamora et al. also evaluated the concurrent validity of the PCS12 and MCS12 in a non-institutionalized population using number of chronic conditions and found similar behavior [37
Several limitations exist with this study. First, despite being nationally representative, our sample is predominantly white, well-educated women; therefore, it is unclear whether the SF-12v2 would perform differently in other patient groups with pain. Secondly, these results cannot be generalized to those who are institutionalized with NCP. Thirdly, a comparison of the PCS12 and the MCS12 to the EQ-5D would have been beneficial to this paper; however, the EQ-5D is no longer administered in MEPS. Next, using number of chronic conditions for concurrent validity was limited by the fact that these conditions were not weighted for the differing severity of chronic diseases. Further, the definition for NCP was defined on meeting the criteria on only one of the four MEPS questions. It could be that patients indicating work limitations did not actually have limitations due to NCP, but rather may have limitations due to other causes. The IOM definition for NCP includes work limitations. This criterion may or may not be representative of patients with NCP, as the IOM notes. For this reason, a sensitivity analysis was performed excluding this criteria from the definition. All results for reliability and validity were similar to the main results. Lastly, recall bias and bias due to missing data could be present in survey data like MEPS.
This study showed that the SF-12v2 was a reliable and valid instrument for measuring HRQoL in patients with NCP. Internal consistency, test-retest reliability, construct (convergent and discriminant) validity, and criterion (concurrent and predictive) validity have been shown to be adequate. In conclusion, the SF-12v2 can be used as a measurement tool to monitor health outcomes in this population. Additionally, HRQoL information obtained from MEPS could be used to inform health status of patients with NCP on a national level.