A Multicenter Assessment of the Adequacy of Cancer Pain Treatment Using the Pain Management Index

Introduction

Pain secondary to bone metastases can be recurrent and debilitating for patients with advanced cancer. Up to three quarters of patients report inadequate pain relief with analgesics,^1–10 and 17%–70% of these patients describe their pain as severe.^{2,3,6–9,11–13} The World Health Organization (WHO) guidelines for the treatment of cancer pain were developed in 1986 (revised in 1996) for the purpose of decreasing the prevalence of inadequately treated pain. ¹⁴ For mild pain, the first step should be a nonopioid medication such as a nonsteroidal anti-inflammatory drug (NSAID) or acetaminophen. If the pain persists or increases to a moderate level, a weak opioid such as codeine is the next step. With severe pain or pain inadequately treated by weak opioids, a strong opioid, e.g., morphine, hydromorphone, oxycodone or fentanyl is recommended. ¹⁴ However, despite the existence of these guidelines and others, undertreatment of cancer pain remains prevalent. ¹⁵

The Pain Management Index (PMI) developed by Cleeland et al. ¹⁶ is a validated method of determining congruence between a patient's reported pain intensity and strength of analgesic prescribed. In a recent review of studies that assessed adequacy of cancer pain treatment, 60% of included studies used the PMI and 80% of these used Cleeland's version. ¹⁶ The adequacy of pain management reported internationally was often suboptimal, with proportions of patients undertreated reaching 81.9% (Fig. 1). There were no studies included that described Canadian cohorts because these data have not been published.

OPEN IN VIEWER

FIG. 1.

A comparison of international studies using the Pain Management Index (PMI) to document undertreatment of cancer pain.

Our primary objective was to use the PMI to determine pain management adequacy in a multicenter cohort of patients with cancer with painful bone metastases referred for outpatient palliative radiotherapy (RT) at three Canadian cancer centers. The secondary objectives were to determine geographic differences in pain severity and PMI between Canadian centers, and to compare our results with similar investigations performed internationally.

Methods and Materials

Data collected prospectively for patients referred to the palliative clinics at the Odette Cancer Centre (OCC) in Toronto, Ontario, the Palliative Radiation Oncology Program (PROP) at the Princess Margaret Hospital (PMH) in Toronto, and the Rapid Access Palliative Radiotherapy Program (RAPRP) at the Cross Cancer Institute (CCI) in Edmonton, Alberta, were reviewed. Inclusion criteria were histologic confirmation of malignancy, radiographic evidence of bone metastases, referral for palliative RT, ability to provide a pain score and analgesic information, cognitively intact, greater than 18 years of age, and English-speaking. Ethics approval was secured from each center.

Data from the OCC were collected from January 1999 to December 2008, while data from the PMH and RAPRP was available for the years 2007–2008. Data extracted included patient age, gender, Karnofsky Performance Status (KPS) and/or Palliative Performance Scale (PPS), primary cancer site, pain score on the Brief Pain Inventory (BPI) or Edmonton Symptom Assessment System (ESAS), and most potent analgesic prescribed at the time of RT consultation.

To determine the PMI, a patient reporting a pain score of zero at the time of assessment in the clinic was assigned a score of 0, mild pain (1-4 on the ESAS/BPI) was assigned a score of 1, moderate pain (5–6) a score of 2, and severe pain (7–10) a score of 3. When a range of scores was described, e.g., 8–10 of 10, the highest rating was used. The analgesic score was calculated based on the potency of medication prescribed by the referring physician. Hence, no analgesic was assigned a score of 0, nonopioid is 1, weak opioid was 2, and strong opioid was 3. For patients using multiple analgesics of varying potency, the most potent analgesic was used for the calculation. ⁴ The PMI is calculated by subtracting the pain score from the analgesic score, and ranges from −3 to +3. Pain management is considered adequate if there is congruence between these two measures (PMI of ≥0), while a negative score represents undermedication. Only data from the initial RT consultation were included and moderate and severe pain was amalgamated for consistency and comparability with other studies.

Descriptive statistics were used for the demographic data. χ² test evaluated the relationship between PMI and the RT center, and PMI and patient-reported pain intensity. Results were considered significant at the 5% level and the tests were two-sided. Statistical Analysis Software (version 9.1, SAS Institute, Cary, NC) was used for the analysis.

Results

Two thousand eleven patients from three specialized Canadian RT clinics were included. Of these, 1144 came from the OCC, 732 from PROP, and 135 from the RAPRP. The results on age and KPS were distributed equally for each center where the mean age of patients was 66.5 and the mean KPS was 60.9 (Table 1). In all centers, the majority of patients presented with a primary diagnosis of lung, breast, or prostate cancer. The incidence of severe pain was significantly correlated with primary cancer site (p = 0.03) when adjusting for the cancer center as the confounder. The majority of patients with breast (57%) and prostate (69%) cancer came from the OCC, and approximately equal numbers of patients with lung cancer were contributed by the OCC (46%) and PROP (49%; Table 1).

The overall proportion of patients with a negative PMI and moderate-to-severe pain was 25.1% and 70.9%, respectively. The incidence of a negative PMI was 31.0% for patients referred to the OCC, 20.0% for the RAPRP, and 16.8% for PROP (p < 0.0001; Fig. 2). The incidence of severe pain was 55.5% for patients seen at the OCC, 48.2% for the RAPRP, and 43.4% for PROP (p < 0.0001). There were no significant differences between centers for the proportion of patients reporting moderate or mild pain (p = 0.35 and p = 0.96, respectively). Internationally, negative PMI scores have been reported between 45%–80% and moderate-to-severe pain between 40%–90% (Fig. 1).

OPEN IN VIEWER

FIG. 2.

Comparison of the incidence of negative Pain Management Index (PMI) and moderate to severe pain in specialized radiotherapy clinics in Canada.

Discussion

The impact of untreated cancer pain can be devastating to the quality of life for patients. The main objective of this study was to determine pain management adequacy in a multicenter cohort of patients with cancer with painful bone metastases referred for outpatient palliative RT at three large Canadian specialized palliative cancer clinics. Differences on undertreatment and severe pain between centres were identified using Cleeland's PMI tool. Cleeland's PMI tool is a simple, widely applicable measure that allows clinicians to make a quick and general comparison of treatment adequacy between different patient populations and settings. Since majority of studies that have used the PMI have incorporated Cleeland's version, our study also used Cleeland's PMI to allow for comparisons within the literature. ¹⁶ However, despite its widespread use, limitations are acknowledged.^17–19 First, it captures a two-dimensional analgesic score that does not account for dose, route of administration, schedule of administration, use of concurrent but less potent medications, nonpharmacologic strategies, or patient adherence to taking the prescribed analgesics. ¹⁷ Second, Cleeland's PMI does not capture a multidimensional pain score. Rather, it provides a snapshot of a patient's worst pain score at one point in time.

Furthermore, it does not account for a patient's tolerance for pain, or for a prescriber's response to a patient's change in pain score. ¹⁸ Third, patient satisfaction is not captured. As a result, a patient with mild pain may have an adequate PMI score with no active treatment suggested; however, this may not be acceptable to the individual patient. ¹⁹ Nonetheless, Cleeland's PMI still remains the most common assessment tool for undertreatment. ¹⁶

Variations in underreporting, poor communication, and compliance with filling analgesic prescriptions may have influenced the results between centers. For example, at the OCC and PROP clinics, the nurse reviews with the patient the medications they have been taking within the last 24 hours. However, at the RAPRP, details on medication history, patient compliance, and counseling on analgesics are discussed with every outpatient by the pharmacist.

Differences in referring health care provider analgesic practices may be another explanation for the differences observed between clinics. For example, variation in knowledge, training, and experience about pain and side effect management for physicians, nurses, and pharmacists likely influenced the community prescribing and/or dispensing of analgesics.^20–26 For example, while the referral patterns were not captured and quantified in this study, generally referrals to each clinic are initiated by medical oncologists. However, palliative care physicians also commonly refer to RAPRP. This may explain why the RAPRP clinic observed relatively less patients with negative PMI scores. At the OCC the incidence of negative PMI scores was higher than the PROP and RAPRP potentially due to the increased frequency of referrals from other specialist physicians such as respirologists and urologists. Other explanations may include that since patients were being referred for palliative RT for pain control, referring physicians may have decided not to subject those patients to the side effects of opioids. Formal assessment practices by referring physicians, differences in follow-up, documentation, knowledge, and access to nonpharmacologic therapies may also explain some of the variation observed.^20–23

Our cohort of patients had a relatively lower incidence of negative PMI compared to international publications on the PMI (Fig. 2). We only compared our results to other international studies that also used the PMI on cancer cohorts. While our cohort of patients was only referred for palliative radiotherapy for uncontrolled metastatic bone pain, the degree of homogeneity is unknown within the other study groups that were compared. While all studies examined cancer cohorts, differences in specific cancer groups such as type or reason for referral, type of cancer, etc., may have attributed to the discrepancy between studies. Differences observed between our results relative to other studies published internationally may be related to variations in the referring institutional practices on addressing pain,^21,27–29 differences in access to palliative care or pain specialists within the referring community, ³⁰ in reimbursement for time spent on symptom management,^20,21 and cultural beliefs of pain. ³¹ For example, in Asian countries negative PMI scores ranged from 45%–70% and moderate-to-severe levels of pain were correspondingly 40%–90% (Fig. 1). These scores are higher when compared with our results (Fig. 2). This may be related to the fears about addiction reported most commonly by physicians from Asia.^15,23,32,33 Additionally, cultural practices from non-Western cultures may attach a spiritual significance to pain. For example, they may expect uncontrolled pain as an inevitable consequence of cancer, or may be suspicious of modern medicine and will not seek conventional treatment at all. ³¹

Conclusions

To our knowledge, this is the first study with the largest sample of patients reporting on undertreated pain using the PMI. This is also the first study reporting on the incidence of negative PMI within the context of a multicenter cohort of patients referred for outpatient palliative RT consultation for metastatic bone pain in Canada. We demonstrate geographic differences between centers for the undertreatment of pain, demonstrating that patients may be still in severe pain when they come for RT consultation.

Future initiatives may include reviewing the differences in pain management using the PMI for patients referred for RT to the nonpalliative RT clinics as compared to dedicated palliative RT clinics. The prescriber's knowledge or viewpoint on analgesics prior to an RT referral, and the referral pattern were not captured, however these may be interesting to consider in future investigations. Developing educational strategies tailored to the general and specific needs of patients and health care providers within these clinics may be a future consideration. Finally, to help improve the quality of care for these patients, implementing a standardized assessment for pain across these clinics should be considered.