Errors in Opioid Prescribing: A Prospective Survey in Cancer Pain

Article Outline

• Abstract
• Introduction
• Methods
• Results
• Discussion
• Conclusions
• Acknowledgments
• Appendix 1. Data Collection Sheet
• References
• Copyright

Abstract 

Context

Cancer pain is debilitating and has multidimensional consequences. It can be treated adequately in up to 90% of patients by following pain management guidelines. Nevertheless, inadequate pain control remains a global problem.

Objectives

We surveyed prescribing patterns in patients referred to our Palliative Medicine Program (PMP) to identify common errors in opioid use.

Methods

Consecutive cancer patients seen by our PMP were prospectively surveyed for the presence of pain and errors in opioid prescribing at the time of initial consultation. Our recommendations to correct and optimize pain management also were recorded.

Results

One hundred eighty-six consecutive cancer patients were screened. One hundred seventeen (63%) had cancer pain, 151 opioid prescribing errors were detected, and 147 different recommendations were made. Most common were failure to order around-the-clock opioids for constant pain, and the failure to treat or prevent opioid side effects. Multiple errors were more common in females, but the sex difference did not reach statistical significance. There was no difference in the errors by pain severity or reason for consultation.

Conclusion

Opioid prescribing errors were common. Females may be at greater risk of multiple errors. A PM consultation program is effective in identifying and correcting a wide variety of opioid prescribing errors.

Key Words: Opioid prescribing errors, cancer pain, palliative medicine

Back to Article Outline

Introduction 

Pain in advanced cancer is complex and is one of the most common and distressing symptoms. It is moderate to severe in 40%–50% of individuals and severe to excruciating in 25%–30%.¹ Patients may experience a combination of several pain syndromes, including somatic, visceral, and neuropathic pain.² Pain has different temporal patterns, including constant, intermittent, and mixed.

Opioids remain the mainstay of treatment for cancer pain. Adequate knowledge of their pharmacology is crucial for prescribing the proper opioid, frequency, dose, and route of administration. However, these choices are complicated by several factors, including difficulty in pain assessment, variable treatment response, and variable individual response.³

Failure to manage cancer pain properly can result from a greater interest in the disease than its consequences. Other factors include 1) limited resources, 2) legal concerns, 3) insufficient medical staff education in pain control, and 4) failure to follow guidelines.⁴ Successful cancer pain management can be achieved in 70%–90% of patients when the World Health Organization (WHO) stepladder analgesic guideline is adopted.⁵, ⁶, ⁷ The goal is effective pain control through the use of a limited number of potent medications. Correct dosing of these medications is critical for success.⁸, ⁹

Errors in pain management have been classified into 1) assessment and documentation, 2) treatment and management, and 3) patient education.¹⁰ The first two are related to professional education. Identifying and understanding these errors may facilitate professional education in proper opioid prescribing and emphasizes the need to follow guidelines to optimize pain management. Strategies for opioid selection, dose timing, and route of administration influence opioid response.¹¹ Physicians frequently fail to implement guidelines, resulting in inadequate analgesia and needless suffering.¹², ¹³ We have previously identified opioid dosing errors and recommended dosing strategies to avoid these errors.⁸ In this study, we prospectively surveyed errors observed during consultations in cancer patients with pain. The objectives were to 1) identify and describe errors in opioid dosing strategies, 2) report their prevalence, and 3) document strategies suggested to correct them.

Back to Article Outline

Methods 

We conducted a prospective survey of opioid prescribing strategies in cancer patients with pain. All new outpatients and inpatients referred to our Palliative Medicine Program (PMP) at The Cleveland Clinic over a period of 80 days in 2004 were evaluated. The study protocol was approved by the Institutional Review Board. A waiver of consent was obtained, as it was an observational study of our practice. This also was a part of a larger study¹⁴ that evaluated the services provided by the PMP.

Patients were seen in consultation in the outpatient clinics, hospital consultation, on our PM acute care inpatient unit, inpatient hospice unit, and in home visits. Those with noncancer diagnoses were excluded. Data collection included patient demographics (age, gender, and race), primary diagnosis, stage of disease, location of consult, referring service, and reasons for consultation. If cancer pain was present, pain severity, prescribed opioids, appropriateness of the current pain regimen, and errors in prior opioid prescribing were documented. A standard data collection tool was used for documentation (Appendix); each patient was assigned an identifying code number. Staff physicians, fellows, and a nurse practitioner gathered the data, which were then entered into a password protected Microsoft Excel^® sheet (Microsoft, Redmond, WA).

Opioid dosing strategies were derived principally from our own practice, but also from various sources, including the Agency for Health Care Policy and Research, European Association of Palliative Care, WHO, and American Pain Society. Deviations from these strategies were considered errors and categorized into four groups: 1) opioid dosing strategy, 2) conversion or rotation, 3) dose titration, and 4) failure to use an adjuvant analgesic. Additional errors could be described in a separate group when applicable (Table 1).

Table 1. Types of Opioid Prescribing Errors

SR=sustained release; CI=continuous infusion; IR=immediate release.

The likelihood ratio Chi-square test was used to determine any significant difference in error distribution between males and females and between older and younger individuals (P<0.05 to reject the null hypothesis of no difference). Binomial tests were used to determine any significant difference in distribution in this group compared with the hypothesized 70% of errors in the whole group (P<0.05 was significant). A separate analysis was done in a subgroup with pain ≥4 (on a scale of 0–10) and in those we consulted on primarily for symptom management.

PMP attending physicians and clinical fellows assessed the current opioid regimen at the time of first consultation in relationship to pain syndrome, severity, temporal pattern, opioid route of administration, and side effects. The final determination about the presence or absence of any errors, and the decision about suggested recommendations were made by the responsible staff attending physician. These were influenced by international practice and guidelines but determined by our unit protocols.⁴, ⁸, ⁹

Back to Article Outline

Results 

We assessed 238 consecutive new patients referred to our PMP over a period of 80 days. Fifty-two had noncancer diagnoses and 186 had cancer (Fig. 1). Patient demographics are provided in Table 2. Seven patients had more than one malignancy. There were no known systematic errors in data collection.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Breakdown of study patients.

Table 2. Patient Characteristics

Characteristic	n=186 (%)
Age, median (range)	63.5 (18–90)
Gender
Male	88 (47)
Female	98 (53)
Race
White	136 (73)
Black	47 (25)
Hispanic	3 (2)
ECOG score
0	13 (8)
1	26 (14)
2	45 (24)
3	61 (33)
4	37 (20)
Common cancer diagnoses
Lung	27 (14)
Breast	16 (8.5)
Colorectal	16 (8.5)
Cancer of unknown primary	15 (8)
Bladder	12 (6.5)
Pancreas	10 (6)
Kidney	10 (6)
Multiple myeloma	9 (5)
Prostate	7 (4)

ECOG=Eastern Cooperative Oncology Group.

Of the 186 patients with cancer, 117 (63%) had cancer pain; 65 (55%) of those were females and 52 (45%) were males. Cancer pain was observed in 62% of those with cancer who were less than 65 years compared with 39% of those aged 65 years and older. This was statistically significant (likelihood ratio Chi square P=0.00021). Pain could not be properly assessed in seven patients because of delirium or because they were actively dying.

Eighty-two (70%) of those with cancer pain had at least one incorrect opioid order. The total number of errors was 151, with a median of 1 (range 1–7). Thirty-seven (45%) had ≥2, 16 (20%) had ≥3, and 7 (9%) had ≥4 errors (Table 3). Although errors were more common in females, the difference did not reach significance (P=0.08). Relatively more females had multiple (>1) errors compared with males (55% vs. 34%, respectively), but this also was not statistically significant (P=0.053).

Table 3. Frequency of Errors

Of the 186 cancer patients, symptom control was the identified reason for consultation in 134 (72%) patients. Sixty-eight percent of this group had cancer pain compared with 46% of those in whom we were consulted for other reasons. There were similar opioid prescribing errors (72% vs. 71%) when we compared consultations for cancer pain with those for other reasons. Sixty-six percent had moderate to severe pain (≥4 on 0–10 scale). The prevalence of opioid prescribing errors was not statistically different compared with those with mild pain (P=0.06).

Opioid dosing strategy issues accounted for most errors (78%), particularly failure to give around-the-clock (ATC) opioids for constant pain. Other common errors were failure to treat opioid side effects, incorrect dosing intervals, and failure to titrate opioid doses (Table 1). Other examples included 1) meperidine use in renal failure, 2) ordering an oral opioid for those unable to take by mouth, 3) using transdermal fentanyl along with parenteral patient-controlled analgesia, and 4) failure to appropriately titrate adjuvant analgesics. Similar patterns were found among those who had moderate to severe pain (Table 4).

Table 4. Opioid Prescribing Errors in Moderate to Severe Pain

Nineteen different dosing strategy recommendations were made by the PM team. A total of 147 recommendations were made with a median of 2 (range 1–6) per patient. The most common recommendations in descending order were 1) starting low-dose continuous opioid infusion, 2) opioid rotation, 3) pre-emptive opioids for incident pain, and 4) management of side effects (Table 5). Other recommendations included discontinuation of one opioid in those receiving two or more opioids, scheduled ATC adjuvant analgesics, patient-controlled analgesia, and nonpharmacological methods for pain control (e.g., radiation therapy).

Table 5. Common Dosing Strategy Recommendations

SR=sustained release; CI=continuous infusion; IR=immediate release.

Back to Article Outline

Discussion 

Pain is one of the major factors affecting health-related quality of life in cancer patients; its optimal management can improve quality of life.¹⁵ Unrelieved pain has a devastating effect not only on patients but also on their families and caregivers.¹⁶ It affects multiple domains in their lives, including the psychological, social, spiritual, and existential, and may precipitate hopelessness and make desire for death more likely. Poorly controlled pain also may increase patient anxiety and concerns related to disease progression and death. During a complex illness, multiple factors can interfere with opioid responses including 1) disease progression, 2) side effects, 3) the type and temporal pattern of the pain syndrome, and 4) individual pharmacokinetic and pharmacodynamic factors.¹¹

The WHO analgesic ladder was developed to provide a guideline for analgesia based on pain severity. At least 10% do not achieve a satisfactory balance between pain control and side effects despite implementing the WHO recommendations and may require further interventions for pain control (e.g., radiation treatment and interventional analgesia). Multiple national, international, and professional pain guidelines exist.⁷, ¹⁷, ¹⁸, ¹⁹, ²⁰ When implemented, they are more effective than standard-practice pain management.⁷ Despite such guidelines, cancer-related pain remains undertreated.¹⁰ Barriers to effective cancer pain management have gained attention and are usually classified into health care provider-related barriers or patient and primary caregiver-related barriers.²¹

Our group previously identified a number of errors in opioid dosing, which we believed were common in clinical practice.⁸ This report has prospectively identified their prevalence in cancer patients and our recommendations in response. Errors were identified and recommendations were suggested in all patients on referral; only those with cancer-related pain were surveyed. The demographic characteristics resemble other reports from our program,²² and hence we believe they represent the population of those dying from advanced cancer. Similar studies, including longitudinal designs, are needed in other settings, such as nursing homes and home care.

In a previous survey performed on inpatients referred to a cancer pain service, 80% underwent changes in opioid selection, route of administration, or both to improve pain relief and diminish the side effects and invasiveness of therapy. This resulted in better pain control and less cognitive impairment and other adverse effects.²³ Our survey showed that 70% of all cancer patients referred had at least one pitfall in opioid prescribing and required recommendations for a new treatment strategy. These recommendations were intended to improve pain control, minimize toxicity, and optimize compliance. It was noteworthy that there was no relationship between prevalence of errors and pain severity.

The prevalence or patterns of opioid prescribing was unaffected by gender (although there was a trend toward more errors in female patients), pain severity, or reason for consultation. This suggests that the problem with erroneous opioid prescribing is extensive in nature and both systems and operator dependent. In part, we believe it reflects deficiencies in physician education about cancer pain control.

Most common in our survey were errors in dosing strategy, predominantly failure to order ATC opioids. This might reflect poor assessment, difficulty in dose titration as an initial strategy, underestimation of pain severity, or inadequate knowledge about opioid prescribing and pharmacology. Failure to recognize different temporal pain patterns (constant vs. intermittent; incident vs. nonincident) can result in poor pain management, and the mislabeling of pain as refractory. Determination of pain type, location, and severity is not always easy, but inadequate assessment is the single most important barrier to effective pain management.¹² Pain assessment requires understanding of its pathophysiology and temporal pattern. For instance, visceral pain, incident somatic pain, and neuropathic pain may be less responsive to opioids.¹¹, ²⁴ Incident pain needs a different dosing strategy; it is important to titrate rescue doses as opposed to increasing the ATC dose.⁹ Dosing for severe incident pain is usually independent of both ATC doses or as needed doses for breakthrough pain.

In a survey evaluating the use of oral morphine by primary care physicians in terminal cancer, the most significant errors were those of dose titration and management of side effects.²⁵ To achieve effective pain control, side effects must be addressed quickly or proactively when appropriate.²⁶ This approach minimizes common opioid side effects, such as neurotoxicity and gastrointestinal toxicity, and improves compliance. Although in our study patients were referred from multiple specialties, failure to titrate opioids and treat opioid side effects also was common (9% and 15%, respectively), hence this issue transcends any single specialty.

Incorrect opioid orders may occur because of poor understanding of opioid pharmacokinetics and pharmacodynamics. Understanding the concept of equianalgesia for opioid rotation and route conversion ratio is essential to pain management and serves to achieve effective pain control with fewer side effects. These calculations should be guided by clinical judgment and the knowledge of equianalgesic tables. Opioids have no analgesic ceiling effect; therefore, it is difficult to define a dose where patients show no response, hence limitations to dose titration should often be related to side effects.²⁷ The development of intolerable adverse effects associated with inadequate analgesia is an indication to consider opioid rotation (with or without an adjuvant analgesic).

The study has some inherent limitations. Our recommendations were derived from published guidelines, validated standards, and recommendations by the Agency for Health Care Policy and Research, European Association of Palliative care, WHO, and American Pain Society. The errors and recommendations were discussed and agreed as our approach to clinical practice by our group prior to our survey. Their interpretation and selection were subject to the discretion of the individual PMP attending staff. Some recommendations or comments on opioid prescribing may have been reported differently by attending staff. Nevertheless, we believe these differences are small. Another caveat is that we did not collect data regarding the status of pain management after the recommendations we made or data regarding their implementation. These should be the subject of further studies. Some recommendations may have not been captured. For example, if we suggested rotation from as-needed morphine to a continuous infusion of fentanyl, we could have recorded opioid rotation but not the simultaneous change to an ATC opioid. This would lead to underestimation of error prevalence. Patients with noncancer pain were not surveyed for the quality of opioid prescribing at the time of consultation, but we believe that a similar pattern might be seen in that population, too.

This survey also reflected the common practice and difficulties of pain management in advanced cancer in a tertiary care academic medical center with many inexperienced physicians (and other professionals) at various stages of training. This is not a reflection of institutional inadequacy in pain control, as we evaluated only those with cancer pain referred to our program, and these results only reflect that population. Furthermore, we critically evaluated opioid dosing strategies according to specialty standards that may not be well-known or taught to those who are not specialized in pain management or PM. This does not change the fact that errors in opioid prescribing according to temporal pain pattern and treatment of side effects were very common. These are important aspects of pain management. These findings should stimulate broader physician education in basic principles of effective opioid use and support further investigation to identify the underlying issues in various practice settings.

Back to Article Outline

Conclusions 

Opioid prescribing errors in cancer pain appear to be both common and complex. They included errors in multiple areas: strategy, conversion, rotation, titration, and the use of adjuvant analgesics. Most commonly, they were failure to prescribe ATC opioids, and manage side effects, specifically constipation prophylaxis and treatment. These errors did not differ with gender, pain severity, or reason for consultation. Females may be at higher risk of multiple errors. Further research is needed in various care settings to identify and eliminate the sources of these errors. A PM consultation effectively identified and corrected multiple complex opioid dosing errors.

Back to Article Outline

Acknowledgments 

The authors thank Mathew Karafa for his statistical assistance.

Back to Article Outline

Appendix 1. Data Collection Sheet 

• 
  • View Large Image
  • Download to PowerPoint

• 
  • View Large Image
  • Download to PowerPoint

Back to Article Outline

References 

1. Ripamonti C, Zecca E, De Conno F. Pharmacological treatment of cancer pain: alternative routes of opioid administration. Tumori. 1998;84:289–300
   • View In Article
   • MEDLINE
2. Davis M, Walsh D. Cancer pain syndromes. Eur J Palliat Care. 2000;7:206–209
   • View In Article
3. Smith HS. Variations in opiod responsiveness. Pain Physician. 2008;11:237–248
   • View In Article
4. Davis MP, Walsh D. Epidemiology of cancer pain and factors influencing poor pain control. Am J Hosp Palliat Care. 2004;21:137–142
   • View In Article
5. Ventafridda V, Tamburini M, Caraceni A, De Conno F, Naldi F. A validation study of the WHO method for cancer pain relief. Cancer. 1987;59:850–856
   • View In Article
   • MEDLINE
   • CrossRef
6. Grond S, Zech D, Schug SA, Lynch J, Lehmann KA. Validation of World Health Organization guidelines for cancer pain relief during the last days and hours of life. J Pain Symptom Manage. 1991;6:411–422
   • View In Article
   • MEDLINE
   • CrossRef
7. Walker VA, Hoskin PJ, Hanks GW, White ID. Evaluation of WHO analgesic guidelines for cancer pain in a hospital-based palliative care unit. J Pain Symptom Manage. 1988;3:145–149
   • View In Article
   • MEDLINE
   • CrossRef
8. Kochhar R, LeGrand SB, Walsh D, et al. Opioids in cancer pain: common dosing errors. Oncology (Williston Park). 2003;17:571–575
   • View In Article
   • MEDLINE
9. Walsh D, Rivera N, Davis MP, Lagman R, LeGrand SB. Strategies for pain management: Cleveland clinic foundation guidelines for opioid dosing for cancer pain. Support Cancer Ther. 2004;1:157–164
   • View In Article
10. Strack PL, Sherwood GD, Adams-McNeill J, Thomas EJ. Identifying and addressing medical errors in pain mismanagement. Jt Comm J Qual Improv. 2001;27:191–199
    • View In Article
    • MEDLINE
11. Mercadante S. Predictive factors and opioid responsiveness in cancer pain. Eur J Cancer. 1998;34:627–631
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (66 KB)
    • CrossRef
12. Von Roenn JH, Cleeland CS, Gonin R, Hatfield AK, Pandya KJ. Physician attitudes and practice in cancer pain management. A survey from the Eastern Cooperative Oncology Group. Ann Intern Med. 1993;119:121–126
    • View In Article
    • MEDLINE
13. Cleeland CS, Gonin R, Hatfield AK, et al. Pain and its treatment in outpatients with metastatic cancer. N Engl J Med. 1994;330:592–596
    • View In Article
    • MEDLINE
    • CrossRef
14. Estfan B, Davis MP, Walsh D, et al. The business of palliative medicine—part 5: service utilization in a comprehensive integrated program. Am J Hosp Palliat Med. 2007;24:211–218
    • View In Article
15. Sandblom G, Carlsson P, Sennfält K, Varenhorst E. A population-based study of pain and quality of life during the year before death in men with prostate cancer. Br J Cancer. 2004;90:1163–1168
    • View In Article
    • MEDLINE
    • CrossRef
16. Coyle N. In their own words: seven advanced cancer patients describe their experience with pain and the use of opioid drugs. J Pain Symptom Manage. 2004;27:300–309
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (190 KB)
    • CrossRef
17. Ressel GW. NIH releases statement on managing pain, depression, and fatigue in cancer. Am Fam Physician. 2003;67:423–424
    • View In Article
    • MEDLINE
18. Hanks GW, Cono F, Cherny N, et al. Expert Working Group of the Research Network of the European Association for Palliative Care Morphine and alternative opioids in cancer pain: the EAPC recommendations. Br J Cancer. 2001;84:587–593
    • View In Article
    • MEDLINE
    • CrossRef
19. Krakowski I, Theobald S, Balp L, et al. FNCLCC Summary version of the standards, opinions and recommendations for the use of analgesia for the treatment of nociceptive pain in adults with cancer (update 2002). Br J Cancer. 2003;89(Suppl 1):S67–S72
    • View In Article
    • CrossRef
20. Benedetti C, Brock C, Cleeland C, et al. National Comprehensive Cancer Network NCCN practice guidelines for cancer pain. Oncology (Williston Park). 2000;4:135–150
    • View In Article
    • MEDLINE
21. Pargeon KL, Hailey BJ. Barriers to effective cancer pain management: a review of the literature. J Pain Symptom Manage. 1999;18:358–368
    • View In Article
22. Walsh D, Donnelly S, Rybicki L. The symptoms of advanced cancer: relationship to age, gender, and performance status in 1,000 patients. Support Care Cancer. 2000;8:175–179
    • View In Article
    • MEDLINE
    • CrossRef
23. Cherny NJ, Chang V, Frager G, et al. Opioid pharmacotherapy in the management of cancer pain: a survey of strategies used by pain physicians for the selection of analgesic drugs and route of administration. Cancer. 1995;76:1283–1293
    • View In Article
    • MEDLINE
    • CrossRef
24. Mercadante S, Maddaloni S, Roccella S, Salvaggio L. Predictive factors in advanced cancer pain treated only by analgesics. Pain. 1992;50:151–155
    • View In Article
    • Abstract
    • Full-Text PDF (637 KB)
    • CrossRef
25. Sanchez Gonzalez R, Ruperez Cordero O, Guerra Merino A, et al. Evaluation of primary care treatment of oncology patients with opiates. Aten Primaria. 1995;16:92–95
    • View In Article
    • MEDLINE
26. Herndon CM, Kalauokalani DA, Cunningham AJ, Jackson KC, Dunteman ED. Anticipating and treating opioid-associated adverse effects. Expert Opin Drug Saf. 2003;2:305–319
    • View In Article
    • MEDLINE
    • CrossRef
27. Hanks GW, Forbes K. Opioid responsiveness. Acta Anaesthesiol Scand. 1997;41:154–158
    • View In Article
    • MEDLINE
    • CrossRef