|
Study 1 Citation: Hwang SS, Chang VT, Fairclough DL, Kasimis B Development of a cancer pain prognostic scale J Pain Symptom Manage. 2002;24(4):366-378
|
Click on the link at left to go to your desired page: Introduction Page 1 Page 2 Page 3 Study 1 Study 2 Study 3 Conclusion Post-Test
Abstract:
This study was conducted to develop a simple bedside tool to predict the
short-term possibility for adequate pain management in cancer patients. 74
cancer patients with severe pain were recruited to participate in this
three-week study. Several assessment tools on personal, pain, distress, and
quality of life characteristics were administered at baseline and weekly
during the implementation of pain management therapies. In addition,
information on treatments was recorded. Each variable was evaluated for its
ability to predict pain relief over the course of treatment to a probability
level of ≥ 0.5 and tested for sensitivity and specificity. Strongest
predictors of pain relief at a level of ≥80% were included at week 1 to
formulate an “intermediate scale” calculation, which was included along with
treatment and type of pain to calculate a Cancer Pain Prognostic Scale. The
resulting score ranged from 0-17 with the higher numbers correlating to pain
relief. The authors suggest that the tool can be used to identify patients
who are less likely to have pain relief.
Cancer Pain Prognostic Scale (CPPS), pain assessment, pain treatment, pain relief
Discussion:
Undertreated pain remains a concern, particularly in cancer patients experiencing complex pain syndromes. While pain staging received recent attention, the ability to predict successful pain relief remained cumbersome. The authors suggest that asking patients about pain and employing an easy-to-use tool to accurately anticipate the success of therapy may remove barriers to a good assessment that can help to solve the problem of undertreatment. This study was undertaken in an effort to develop such a clinical tool.
Previous work completed to develop a prognostic tool separated patients into two primary categories for pain control: good prognosis and poor prognosis. Characteristics are shown in Figure 1.
Figure 1. Characteristics of patients with good versus poor prognosis for successful pain control according to the Edmonton Staging System (ESS)
| Prognosis | Patient Characteristics |
|
Good (Stage I)
|
Visceral or somatic categories of pain;
nonincidental pain; bone or soft tissue pain |
|
Poor (Stage II)
|
Neuropathic pain, mixed pain, or unknown type of
pain; nonincidental pain |
The original ESS included an assessment of cognitive function and opioid dose. These domains were not found to be predictive. While the research to test this information suggested the ability to determine pain management prognosis problems of subjectivity, specificity, and difficulty in conducting testing in a standardized fashion that led to agreement in results between investigators remained unclear. For instance, presence of substance abuse can range from a distant history to current and severe abuse, a difference not picked up in the ESS, but subjectively likely to affect pain management potential. Stage II patients (poor prognosis) were often able to achieve satisfactory pain control in 50% of cases.
In this study, predictive variables included characteristics of pain, quality of life measures, symptoms of distress, and patient characteristics. The model used in the development of the prognostic tool followed guidelines that suggested the inclusion of patient history, physical examination, and other simple tests that could suggest a course of therapy or predict outcomes. The authors also applied standards of prediction rules that included blinded assessment outcomes and predictors and outcomes definitions.
Pain management strategies were initiated on enrollment according to the AHCPR guidelines. Follow-up assessment was completed at least weekly for three weeks with adjustment in medications and changes in pain-related variables. Differences between pain variables from baseline and for each weekly assessment included pain relief and severity of average pain and worst pain as demonstrated by the Brief Pain Inventory (BPI). Variables were tested for prediction capacity and combined to develop a multidimensional model of pain relief prediction. This point system was then tested for positive and negative predictive values as well as sensitivity and specificity to ≥80% pain relief prediction at a level of ≥0.5. The resulting Cancer Pain Prognostic Scale (CPPS) prediction equation included worst pain severity by BPI, emotional well-being by FACT-G, daily opioid dose of >60 mg morphine by mouth, and the presence of mixed pain.
Results:
The resulting possible values range between 0-17 points. Groups were stratified to yield high scores with high possibility of pain relief within one to two weeks at 13-17 points; intermittent scores of 7-12; and low scores of 1-6. The authors rated participants according to the ESS scale and found a significant difference from the CPPS that they attributed to the wider range of possible outcomes (ESS with Stages I and II compared to a graded range in the CPPS).
While the CPPS scale was useful in predicting pain relief at up to two weeks, predicting variables at week three were different from those identified at weeks one and two. Likewise, the ESS scale was poorly predictive beyond the second week of pain treatment.
Summary of Study 1
|
The Edmonton Staging System (ESS) was the only previously available prognostic scale for cancer pain management. The ESS had several limitations including its validation in hospice patients which may not apply in other cancer patients, its use of only two stages (good and poor) which classified many patients into a poor prognosis for pain relief who were able to achieve satisfactory pain management by the third week, and the lack of inclusion of disease, physical distress, psychological issues, symptoms, disease extent, patient barriers, age, and responsiveness to opioid therapy as variables. The authors also suggest that variables that predict outcomes in patients at weeks one and two may no longer be valid at week three because of insensitivity to treatable pain at that point. In this study the researchers determined that quality of life, social situation, symptom distress, and pain characteristics were the best predictors of successful pain management. Interestingly, the presence of neuropathic pain was not independently predictive in the development of the CPPS model. The researchers chose to categorize pain as nociceptive (injury-stimulated pain), bone, breakthrough, neuropathic, and mixed pain. The authors suggested that this categorization could have influenced results because nociceptive pain is a predictor for slow pain relief. Important predictive variables included worst pain severity on the BPI evaluation, emotional well-being on the FACT-G evaluation, and opioid dose. The authors suggested that opioid dose was an important predictor probably because patients maintained on lower doses could benefit from dose increases while patients already on high doses are more likely to have a more severe pain problem. The authors also suggested that substance abuse and general symptom distress were not significant factors in predicting outcomes. Pain relief at week one was best predicted by worst pain severity (BPI) and emotional well-being (FACT-G) measures. Additional variables were added to predict benefit at week two. The calculations for weeks one and two did not apply to week three as different variables appeared to be more predictive. The authors suggested that it would be more difficult to develop an index to predict pain relief success in longer periods of time. Limiting factors in this study included the number of patients and inclusion of a small subgroup that may have been predisposed to poor outcomes. A larger number of patients would be required to determine other variables such as cancer site and pain site. Also, patients enrolled in the study started with a pain score of ≥4 (BPI) and the chosen outcome was set at 80%. Changes in these criteria may have changed the resulting CPPS equation. Additional studies will be needed to address these limitations and further identify variables that can be used to refine a predictive equation for pain management outcomes at an acceptable level. Authors speculated about the potential use of the CPPS or other refined tool suggesting an algorithm that patients showing a high probability of pain relief within two weeks could be treated according to standardized guidelines while patients with a low probability of pain relief may require individualized pain consultation and more intensive management.
|