BACKGROUND/RATIONALE:
The Veterans Administration (VA) faces a substantial challenge in trying to improve end of life care. Foremost among the gaps in research on end-of-life care is inadequate knowledge about how to ensure pain relief. Studies show that pain is a major symptom for patients with a variety of serious chronic conditions. Drawing upon different clinical paradigms for the evaluation and treatment of pain, this study focuses on improving measurement and interpretation of routine pain screening in ambulatory VA patients as well as veterans with CHF as an important step to improving end-of-life care.
OBJECTIVE(S):
Our specific objectives include to evaluate: 1) the reliability and validity of pain as the 5th vital sign by comparing routine pain screening to a) the same numeric scale applied under research conditions, b) other validated measures of pain , and to 2) the association of routine pain screening in general medical illness and CHF with a) unmet need, satisfaction, and provider detection following the treatment visit and b) depression, other symptoms, and co-morbid conditions. 3) the association between routinely detected pain and the quality of outpatient pain management
METHODS:
The study is a cross sectional corhot study in a variety of VA-only outpatient settings (hospital-based, large outpatient multi-specialty, and community-based) at the VA Greater Los Angeles (GLA) and Long Beach (LB) Healthcare Systems. A random sample stratified by self-reported health of cognitively intact veterans who were attending primary care, women's health, cardiology, and oncology patients were surveyed after the visit. Three surveys were used to capture patient, nurse, and clinician perspectives. Chart reviews of encounter data were also conducted. Surveys included a baseline survey of physicians and nurses about pain attitudes and knowledge that evaluated attitudes and knowledge that may influence pain assessment and management practices, drawn from behavioral science theoretical frameworks. Patients as well as physicians were surveyed following encounters, and chart reviews were conducted retrospectively.
FINDINGS/RESULTS:
A total of 6138 patients were screened for eligibility. Of those, 2265 had not had their vital signs taken; 436 were excluded due to good health; 942 had not visited or were not planning to visit a provider that day; 310 were not visiting participating clinics; 103 were visiting providers who did not agree to participate; 51 had cognitive impairment; 49 were hearing impaired; 61 were ineligible on the basis of previous participation; 120 were ineligible on the basis of interviewer judgment (e.g. aggressive behavior); and 862 refused the screener. Among patients approached, 939 were eligible and 650 (69.2%) completed the interview, of whom 646 had their charts reviewed. Of those,19 (2.9%) were missing pain ratings. Of providers, 208/280 (74%) of physicians participated in baseline surveys, as did 146 / 159 (92%) of nurses. 475 / 637 (75%) attempted post visit physician surveys were returned.
In preliminary results, in 51% of encounters, nurses asked about pain without using the 0-10 NRS and that was associated with underestimation of pain (X2 =64.04, p <.001). We found moderate agreement between the 0-10 nurse-NRS and both the research-NRS (r = 0.627) and BPI-severity scale (r = .613 for 24 hour and r = .588 past week version). However, agreement between the research-NRS and BPI was good (24 hours r = .870 and BPI-severity/last week r = .840), suggesting that inaccuracy was related to measurement process, rather than the instrument itself. We have evaluated patient and nurse-related factors associated with variation in pain measurement and the factors associated with under and overestimation on the part of nurses include poorer health status, PTSD, not using the numeric scale (nurse pain underestimation) and use of the numeric scale, professional experience, and lower confidence in pain management (nurse pain overestimation). In collaboration with investigators in Indiana, we have developed a shortened version of the brief pain inventory that may be more useful for screening - that most efficiently identifies variation in intensity as well as interference, but only includes 3 items. Two analyses of pain focused on CHF using the total pain model (physical, social, emotional, and spiritual domains) found pain in about 2/3rd of veterans with CHF) confirmed the independent importance of social and total symptom burden factors - stressing the need for multidisciplinary approaches in serious chronic illness. Preliminary analysis of provider surveys using vignettes found that more highly trained, experienced, and self-confident nursing staff were more likely to provide appropriate counseling, and more appropriate physician management was associated with greater knowledge and skills in the case of prescribing behaviors, and with female gender, the endorsement of the idea that time limits pain assessment, and confidence in non-cancer pain guidelines.
IMPACT:
This study will provide important information to guide implementation of pain as the 5th vital sign throughout the VA healthcare system nationwide (nurse and physician training, and EMR-based strategies for improving collaborative pain management). It will also provide valuable information for other healthcare systems and accrediting organizations implementing routine pain measurement. In addition, it will contribute to a better understanding of the link between routine measurement and basic adherence to recommended pain practices. This study will shed light on the significance of pain in patients with general medical illness and serious cardiac conditions nearing the end of life in which these symptoms are poorly understood. Better understanding of the relationship of pain to other symptoms will inform the development of improved clinical strategies to address pain.
PUBLICATIONS:
Journal Articles
- Lorenz KA, Shugarman LR, Lynn J. Health care policy issues in end-of-life care. Journal of Palliative Medicine. 2006; 9(3): 731-48.
