Background In response to the need to identify measures of the quality of pain management that could be widely implemented in the VA, the VA Performance Measures Workgroup endorsed a pilot of several cancer pain performance measures as a first step.
This pilot aimed to demonstrate the feasibility and methods that could be used for a national evaluation of cancer pain and its management in an appropriate cohort.
Four cancer pain quality indicators (modified from the ACOVE / ASSIST pain QI sets) were identified to include in a VA OQP special chart abstraction for 100% of veterans nationwide diagnosed with colorectal cancer (N=2,896) between 10/1/03 and 03/31/06. We accessed national VA pharmaceutical data on the same cohort. We conducted bivariate comparisons of patient-level pain scores, individual change in pain, and pain medication use by category of pain score (0 [none]; 1-3 [mild]; 4-7 [moderate]; 8-10 [severe]) and by stage of disease. We then calculated pass scores for the four QIs by evaluating the proportion of visits at which patients were eligible and then "passed" each of these four QIs, based on a positive finding for the "THEN" component of the QI. We evaluated these scores as rates, considered different methods for operationalizing the indicators, and evaluated the potential impact of different methods on final pass scores.
Most patients experienced zero pain at either visit, with significantly fewer patients experiencing any pain (pain score >0) at the second outpatient visit compared with at the first. Overall, 99% of patients were on some form of possible pain medication - either opioids, topical analgesics, antidepressants, or anticonvulsants - within this time window, and 98% were on more than one such drug. The most common pain medication was opioids (96% of all patients) and topical analgesics (7% of cohort) were the least common. Pain was less common than expected in this cohort, but among those experiencing pain, breakthrough pain and severe pain at initial visit were managed appropriately over 80% of the time, while sustained or worsening pain and appropriate bowel prophylaxis for opioids were appropriately managed approximately half of the time.
These results demonstrate the feasibility of collecting abstracted chart review data as well as administrative data to evaluate quality of care for pain management, and can be used for future studies to guide selection of appropriate populations and collection of data from appropriate sources. Our analysis provides a basis for methods of operationalizing quality measures, and when done in a larger cohort can help lead to improvements in patient care and health outcomes.
External Links for this Project
None at this time.