Abstract
Purpose
We tested the hypothesis that prescription coverage affects the prescribing of long-acting opiates to indigent inner city minority patients with cancer pain.
Materials and Methods
We conducted a chart review of 360 patients treated in the Oncology Practice at UMDNJ-University Hospital, who were prescribed opiate pain medications. Half the patients were Charity Care or Self Pay (CC/SP), without the benefit of prescription coverage, and half had Medicaid, with unlimited prescription coverage. We evaluated patients discharged from a hospitalization, who had three subsequent outpatient follow up visits. We compared demographics, pain intensity, the type and dose of opiates, adherence to prescribed pain regimen, unscheduled Emergency Department (ED) visits and unscheduled hospitalizations.
Results
There was a significantly greater use of long-acting opiates in the Medicaid group than in the CC/SP group. The Medicaid group had significantly more African American patients and a greater rate of smoking and substance use and the CC/SP group disproportionately more Hispanic and Asian patients and less smoking and substance use. Hispanic and Asian patients were less likely to have long-acting opiates prescribed to them. Pain levels and adherence were equivalent in both groups and were not affected by any of these variables except stage of disease, which was equally distributed in the two groups.
Conclusion
Appropriate use of long-acting opiates for equivalent levels of cancer pain are influenced only by the availability of prescription coverage. The group without prescription coverage and receiving fewer long-acting opiates had disproportionately more Hispanic and Asian patients.
Introduction
Two thirds to three quarters of cancer patients experience pain that requires analgesia (1-3). Cancer pain is complicated by other symptoms that must be considered using evidence-based standards in screening, assessment and treatment (3-5). Treatment must adopt a comprehensive approach that includes nonopiate analgesics, opiates, radiotherapy, psychosocial intervention, appropriate follow up and attention to associated co-morbidities, including opiate-induced constipation (4, 5). Nevertheless, the management of cancer pain in practice varies widely, where the burden of full diagnostic assessment may exceed potential benefits and requires up front symptom relief (4).
Assessment of cancer pain is an equally complicated and a variably-practiced endeavor. Pain scales and assessment instruments have been constructed and validated (6-8) and standards for the treatment of cancer pain have been generated and are in practice (9). Despite these instruments, most pain assessment is done without much thought by many practitioners and standards are often not included in the decision to assess pain (10, 11). One approach however, that of applying an analgesic protocol, was helpful in pain management (12, 13).
Initially, cancer pain may be treated with acetaminophen or nonsteroidal anti-inflammatory agents but opiates should be considered quickly (4). While use of immediate release morphine may be adequate initially, the management of moderate to severe chronic pain should involve a relatively quick conversion to sustained release opiates (14, 15). The use of transdermal Duragesic (fentanyl) and slow release morphine is safe and has become standard therapy for cancer pain (16). Patients receiving long-acting opiate formulations for noncancer pain showed improved treatment responses, a better perception of quality of life, improved focus on daily activities, less focus on pain, better pain regimen adherence and reduced pain-related anxieties (17).
Multiple factors affect the choices for managing pain by physicians, including issues involving the complex nature of cancer pain and its etiology, associated syndromes, associated co-morbidities (3), heritable factors (7), history of substance use, physiologic, pharmacologic, psychosocial factors and gender- and age-related factors (18-20). Physician-related factors influence pain management decisions, which include training biases (21), physician ethnic origin (22), specialty training and the nature and location of the medical facility (23), among others. Patient ethnicity and societal effects significantly influence pain management (1, 2, 24). Ethnic minorities are three times more likely to receive inadequate pain regimens than other patients (1), with approximately two thirds of minority patients not receiving sufficient pain control as compared to half of non-minority patients (2, 23). Specifically, 74% of Hispanic patients and 59% of African American patients did not receive adequate analgesia in one study (2). We have previously demonstrated, in a pilot study, that indigent patients in an urban university hospital were prescribed appropriate long-acting analgesics for the treatment of cancer pain according to whether they had prescription coverage (25). Here, we followed up these preliminary observations in a large scale, definitive study to determine whether having a prescription plan affects the appropriate prescribing of long-acting opiates to indigent minority patients with cancer pain in an academic urban setting. Our data confirm that prescription benefits dictate the use of long acting opiates in these patients.
Methods
Study design, patients and data collection
We planned to test our hypothesis that uninsured indigent cancer patients are prescribed long acting opiate analgesics at a lower rate than patients with unlimited prescription coverage with no deductible out of pocket costs using a retrospective chart review. We reviewed 180 charts of Charity Care/Self Pay (CC/SP) patients and 180 charts of Medicaid patients who were followed in Medical Oncology Outpatient Practice and who met the criteria of receiving treatment for pain, having had a hospitalization and three outpatient follow up visits. There were no time limitations. The patients' payor status was determined by inspecting the University Hospital electronic medical records documentation. New Jersey State Charity Care status is awarded to patients by University Hospital after review of records to demonstrate lack of financial resources and of health insurance coverage. Charity Care entitles patients to free medical care and treatment as inpatients or outpatients in our clinics. These patients have no prescription coverage and are responsible for paying for their own outpatient medications. We included the Self Pay category of patients who are also indigent, uninsured and without prescription coverage but do not qualify for Charity Care due to a variety of issues, including United States residency status. After IRB approval, a total of 360 charts of patients who met the above criteria were selected sequentially from 7/5/2000-8/7/2008 from Medical Oncology Outpatient Department appointment lists printed out randomly.
Data were collected on study sheets that requested a deidentified sequential patient number, age, gender, ethnicity or race, as identified by the medical record, payor status, smoking, alcohol or substance use as yes/no answers, disease, stage and current therapy. The study sheets collected the date of discharge from hospitalization, level of pain reported from the traditional pain scale of 1-10 (8) and analgesic medications and doses prescribed at hospital discharge. The data also included the dates of the three subsequent clinic visits, pain level, pain medications, patient adherence to the prescribed regimen at the time of the visit and reasons for non-adherence. Any Emergency Department visits or unscheduled readmissions and the reasons were also recorded. Our primary endpoints were 1) difference in pain control between two payer groups on day 0, at 1st, 2nd, 3rd visit, 2) differences between two payer groups in the change in pain control for each group between day 0 and 1st, 2nd, 3rd visit or a scheduled admission for chemotherapy, 3) differences in pain regimen between two payer groups on day 0 (use of long acting opiates: 1 vs. 0), 4) differences in adherence to pain regimen at 1st, 2nd, 3rd visit or scheduled admission for chemotherapy between two payer groups, 5) correlation between adherence to pain regimen (change in regimen) and change in pain control in each group and 6) presence of confounding factors, such as differences in age and gender on the above endpoints. Secondary endpoints were 1) differences in the number of unscheduled ER visits between two payer groups and 2) differences in number of unscheduled hospitalizations for pain control or failure to thrive between two payer groups.
Statistical analysis
Using the data from our pilot study of 40 patients (25), we estimated that we would need to include 360 subjects divided equally between Medicaid and CC/SP in this definitive study to allow us to test for statistical significance of the logistic regression coefficients at the 5% level of significance with 80% power (26). The estimates were based on an odds-ratio of 2.5 when comparing severe pain to no pain in the Medicaid vs. CC/SP patients (25).
A widely used rule for judging the adequacy of the sample size for fitting a model is 10 events (lower frequency outcomes) per parameter. With 360 patients, we expected approximately 150 events (long acting prescriptions), giving the potential of a 15 parameter logistic regression model. Based upon the results of our pilot study, this was judged to be sufficient (27). Chi square analysis was used to study the relationship between reported pain levels and demographics such as age, gender and ethnicity. Repeated measures ANOVA was used to study the relationship between reported pain and disease stage and sequence number of the visit at which the pain report was recorded, hereafter referred to as the appointment number. Univariate and simpler bivariate analyses were carried out using Chi square and Mann-Whitney U tests.
Results
Patient characteristics
Age and gender distributions were statistically similar. The mean ages of CC/SP and Medicaid patients were 48.6 ± 13.9 and 48.6 ± 11.1, respectively, and both groups had 53.9% males and 46.1% females. The ethnic and racial distribution (p = 0.000), smoking (p = 0.000) and substance abuse rates (p = 0.008) were significantly different between the two payor groups by univariate chi square analysis (Table 1). The CC/SP group had disproportionately fewer African Americans and the Medicaid group had disproportionately more African Americans than the average of the two groups combined (37.2% vs. 57.8%) (Table 1). Conversely, the CC/SP group had a significantly higher representation of Hispanics and the Medicaid group has fewer Hispanics than predicted (41.1% vs. 25.6%). Caucasians were relatively evenly represented in the two groups at 13.9% compared to 11.1% while CC/SP group had more Asians at 7.2% and the Medicaid group had fewer Asians, at 2.8% (p = 0.000). There was a significantly lower smoking rate (24.0% vs. 42.1%, p = 0.000) and substance abuse rate (4.9% vs. 13.4%, p = 0.008) in the CC/SP group than in the Medicaid group by Chi square analysis. Ethanol use was not statistically significant between the two groups (Table 1).
Table 1. Characteristics of the study patients from the two payer categories.
| Charity Care/Self Pay | Medicaid | p | test | |
|---|---|---|---|---|
|
|
|
|
|
|
| Mean age ( ± SD) | 48.6 ± 13.9 | 48.6 ± 11.1 | 0.627 | Mann-Whitney |
| Gender | ||||
| Male | 97 (53.9%) | 97 (53.9%) | ||
| Female | 83 (46.1%) | 83 (46.1%) | ||
|
|
|
|||
| Total | 180 | 180 | 1.000 | Chi square |
| Ethnicity | ||||
| African American | 67 (37.2%) | 104 (57.8%) | ||
| Hispanic | 74 (41.1%) | 46 (25.6%) | ||
| Caucasian | 25 (13.9%) | 20 (11.1%) | ||
| Asian | 13 (7.2%) | 5 (2.8%) | ||
| Not specified | 1 (0.6%) | 5 (2.8%) | 0.000 | Chi square |
| Behavioral history (of those reporting) | ||||
| Smoking | 42/175 (24.0%) | 75/178 (42.1%) | 0.000 | Chi square |
| Ethanol use | 30/173 (17.3%) | 41/178 (23.0%) | 0.184 | Chi square |
| Substance use | 8/163 (4.9%) | 23/172 (13.4%) | 0.008 | Chi square |
The same six tumor types, colorectal, head and neck, lung and breast carcinoma and sarcoma, were the most frequently represented in both groups (Table 2). However, when comparing the frequency of tumors that occurred in at least 5 patients in both arms, head and neck, lung, breast and cervical carcinomas, sarcomas, and multiple myelomas were more prevalent in the Medicaid group while colorectal and pancreatic carcinomas, germ cell tumors and lymphomas were more common in the CC/SP group (p = 0.017, Chi square). In order to diminish the potential implications of this unbalance as a confounding factor, we stratified the tumor types by stage, which is one of the most significant factors in the odds ratio for occurrence of pain in the clinical setting of malignancies (28, 29). Indeed, the data demonstrate that patients with higher tumor stages were more highly represented in both payor groups receiving analgesic therapy, but the differences in tumor stage distributions were not statistically significant between the two groups (Table 2). The percentages of patients receiving therapy for their disease during the sampling period were also similar and the time to follow up between discharge and the subsequent clinic visits were not statistically significant by a non-parametric test (Table 2).
Table 2. Cancer type distribution, treatment and follow up schedule.
| Charity Care/Self Pay | Medicaid | p | test | |
|---|---|---|---|---|
|
|
|
|
|
|
| Cancer type distribution | ||||
| colorectal | 23 | 19 | ||
| head and neck | 22 | 37 | ||
| lymphoma | 22 | 5 | ||
| lung | 20 | 23 | ||
| breast | 14 | 21 | ||
| Sarcoma | 11 | 17 | ||
| germ cell | 9 | 4 | ||
| pancreas | 7 | 4 | ||
| stomach | 7 | 7 | ||
| cervix | 6 | 9 | ||
| multiple myeloma | 5 | 9 | 0.017 | Chi square |
| unknown primary | 5 | 2 | ||
| CNS | 5 | 2 | ||
| esophagus | 2 | 5 | ||
| anus | 2 | 5 | ||
| ovary | 4 | |||
| liver | 3 | |||
| leukemia | 3 | 2 | ||
| uterus | 3 | 2 | ||
| prostate | 2 | 1 | ||
| neurendocrine tumor | 2 | |||
| kidney | 1 | 3 | ||
| biliary tract | 1 | 2 | ||
| urogenital tract | 1 | 1 | ||
| Stage (of those documented) | ||||
| I | 8 | 8 | ||
| II | 15 | 8 | ||
| III | 30 | 22 | ||
| IV | 67 | 68 | 0.473 | Chi square |
| Patients receiving therapy during sampling period | ||||
| Chemotherapy, molecular target therapy, hormone blockade and/or bisphosphonates (no radiation) | 16 | 15 | ||
| Radiation therapy (with or without concurrent chemotherapy) | 1 | 3 | 0.627 | Chi square |
| Time to follow up visit (days) | ||||
| To 1st follow up | 30.9±46.4 | 27.4 ± 38.0 | 0.620 | Mann-Whitney U |
| To 2nd follow up | 66.5± 64.2 | 65.9 ± 56.3 | 0.697 | Mann-Whitney U |
| To 3rd follow up | 103.2± 72.1 | 110.0 ± 73.7 | 0.114 | Mann-Whitney U |
Pain level
The frequency distribution of patients reporting a specific pain level did not differ significantly between the two groups by Chi square analysis (p = 0.377) (Table 3). The four pain assessments for each patient were averaged to reveal the single value reported in the table. The median pain levels in the two groups did not differ either, when compared directly (p = 0.161). Median pain levels reported were similar in the two genders (p = 0.505) and the different ethnicities (p = 0.374) recorded by Mann-Whitney and Chi square analysis (Table 3). Similarly, none of the variables of age, gender and ethnicity were significant predictors of pain level in Chi square analyses (Table 3). The lack of influence of ethnicity on pain level is particularly noteworthy, as it removes it as a confounding factor in the setting of the ethnic imbalance between the two payor groups.
Table 3. Pain level.
| Charity Care/Self Pay | Medicaid | p | test | |||
|---|---|---|---|---|---|---|
|
|
|
|
|
|||
| Pain level* | (number of patients) | (number of patients) | ||||
| 1 | 21 | 14 | ||||
| 2 | 30 | 31 | ||||
| 3 | 30 | 32 | ||||
| 4 | 33 | 29 | ||||
| 5 | 34 | 25 | ||||
| 6 | 14 | 19 | ||||
| 7 | 9 | 18 | ||||
| 8 | 7 | 12 | ||||
|
|
||||||
| Total | 178 | 180 | 0.377 | Chi square | ||
| Median pain level [Quartiles] | 4 [0,6] | 4 [0,6] | 0.161 | Mann-Whitney U | ||
| Median pain level by gender [Quartiles] | ||||||
| Male | 4 [0,7] | |||||
| Female | 4 [0,6] | 0.505 | Mann-Whitney U | |||
| Pain Level** | African-American | Caucasian | Hispanic | |||
|
|
||||||
| 2 or less | 30 | 10 | 30 | |||
| 3 | 35 | 7 | 20 | |||
| 4 | 29 | 8 | 27 | |||
| 5 | 34 | 5 | 13 | |||
| 6 | 13 | 7 | 15 | |||
| 7 or more | 30 | 8 | 15 | |||
|
|
||||||
| Total | 171 | 45 | 120 | 0.374 | Chi Square | |
| Pain Level** | Below 40 | 40-49 | 50-59 | 60.and -59 above | ||
|
|
||||||
| 2 or less | 18 | 12 | 27 | 20 | ||
| 3 | 19 | 11 | 25 | 10 | ||
| 4 | 14 | 22 | 18 | 13 | ||
| 5 | 9 | 14 | 22 | 10 | ||
| 6 | 10 | 10 | 12 | 9 | ||
| 7 or more | 12 | 16 | 21 | 6 | ||
|
|
||||||
| Total | 82 | 85 | 125 | -68 | 0.374 | Chi Square |
| Mean pain levels at stage vs. each appointment | ||||||
| Discharge | Appt 1 | Appt 2 | Appt 3 | |||
|
|
|
|
|
|||
| Stage | ||||||
|
|
||||||
| I | 6.13 | 2.38 | 2.06 | 1.97 | ||
| II | 6.17 | 2.57 | 3.96 | 3.22 | ||
| III | 5.15 | 3.00 | 2.88 | 2.58 | ||
| IV | 5.77 | 3.56 | 3.48 | 3.44 | ||
| Stage by appointment interaction | 0.551 | |||||
| Equal stage marginal means | 0.136 | |||||
| Equal appointment marginal means | 0.000 | multivariate ANOVA witd repeat measures for unbalanced designs | ||||
(1439 observations across 4 appointments averaged by patient);
(Asian patients dropped due to too few observations)
For pain levels at stage vs. each appointment, multivariate analysis of variance (ANOVA) with repeated measures for unbalanced designs was carried out. Stage did not affect the number of appointments attended (stage by appointment interaction, p = 0.551). The mean pain levels did not change significantly with subsequent appointment number in patients at the same stage (test for equal stage marginal means, p = 0.136). However, the average pain level did increase significantly with stage at each appointment, albeit not linearly (test for equal appointment marginal means, p = 0.000), in congruence with the increased numbers of patients being treated for pain with progressive stage (Table 2). There were no significant differences in mean pain levels at each visit between the payor groups either (not shown).
CC/SP patients visited the University Hospital emergency department at significantly lower rates than Medicaid patients (14.4% vs.23.9% of the patients, p = 0.023 with 54 vs. 71 total ER visits for the two groups (p = 0.000). Of the reasons for the visits 29 vs. 56 of the visits in the two payor groups were for pain (Table 4). However, a comparison of all the reasons for ER visits that were reported at least 5 times in each group did not yield a statistically significant difference (p = 0.737). While, fewer CC/SP patients had unscheduled readmissions to University Hospital than Medicaid patients (29.4 vs. 37.6%, the difference also did not reach statistical significance (p = 0.100) (Supplementary Table 1). CC/SP patients had 89 readmissions vs. 114 in the Medicaid group (p = 0.079), but unlike ER visits, the causes of readmission were for a variety of medical problems as well as pain, and the differences between the most common reasons were not statistically significant (p = 0.463) (Supplementary Table 1).
Table 4. Emergency Department Visits.
| Charity Care/Self Pay | Medicaid | p | test | |
|---|---|---|---|---|
|
|
|
|
|
|
| Patients with emergency department visits | 26/180 (14.4%) | 43/180 (23.9%) | 0.023 | Chi square |
| Total ER visits | 54 (0.26/patient) | 71 (0.47/patient) | 0.021 | Mann-Whitney |
| Causes of ER visits (some visits had multiple causes) | ||||
| -Pain (including abdomen, back, breast, chest and acute myocardial syndrome, extremities, eye, face, flank, groin, head, hip, mouth, penis, rectum, vagina and unspecified pain or running out of pain medication) | 29 | 56 | ||
| -Nausea/vomiting | 7 | 10 | ||
| -Weakness, dehydration, falls, syncope, presyncope, gait issues | 6 | 8 | 0.737 | Chi square |
| -Infection, leucopenia or fever | 5 | 3 | ||
| -Breathing issues | 2 | 5 | ||
| -Malfunctions of tubes, devices | 2 | 5 | ||
| -Bleeding | 2 | 3 | ||
| -Neck swelling | 2 | |||
| -Diarrhea/constipation | 2 | 4 | ||
| -Lab abnormalities | 1 | 2 | ||
| -Drainage or pus | 2 | |||
| -Allergic rxn | 1 | |||
| -Bone in throat | 1 | |||
| -Hypertension | 1 | |||
| -Cough | 1 | |||
| -Anemia | 1 | |||
| -Visual disturbance | 1 | |||
| -Palpitations | 1 |
Use of long-acting opiates
The number of CC/SP patients who were prescribed long acting opiates at any time during the four encounters was significantly lower at 55 (30.6%) than the number of Medicaid patients at 98 (54.4%) (p = 0.000 Table 5). This difference was due to the significantly lower rate of use of all three long acting opiates studied, Duragesic patches (p = 0.000), Oxycontin (p = 0.002) and MS Contin (p = 0.051), by CC/SP patients than by Medicaid patients. Comparisons in rates of use were made between the number of patients prescribed each drug at any time and the number of patients who were never prescribed long acting opiates (125 for CC/SP and 82 for Medicaid patients). The median doses of the three drugs were not statistically different between the two payor groups (Table 5). Methadone was rarely and inconsistently prescribed and was not included in the analysis. In line with the ethnic weighting of the two payor groups, Hispanic and Asian patients were prescribed long acting opiates at a lower rate than that expected by the distribution of use in the entire sample (p = 0.027) (Table 5). African Americans and Caucasians, on the other hand, were prescribed long acting opiates at a higher rate than the average rate of all the patients sampled. There were no differences in long acting opiate use between the genders (p = 0.448) and the number of patients adhering to the regimen at each follow up visit did not differ between the two payor groups (p = 0.94). There were no differences in the change from use of long term opiates from hospital discharge to clinic between the two groups (not at discharge -> yes in clinic 22 vs. 26, yes at discharge -> no in clinic 9 vs. 9). There was a significantly greater use of long acting opiates with progressive stage in all patients (p = 0.006), SP/CC patients (p = 0.000) and Medicaid patients (p = 0.014) (Table 5), congruent with the increased levels of pain reported with progressive stage (Table 3).
Table 5. Use of long-acting opiates.
| Charity Care/Self Pay | Medicaid | p | test | |
|---|---|---|---|---|
|
|
|
|
|
|
| Patients with any use of long acting opiates | 55/180 (30.6%) | 98/180 (54.4%) | 0.000 | Chi square |
| Patients using specific long-acting opiates at any time | ||||
| Duragesic | 28 | 57 | 0.000 | Chi square |
| Oxycontin | 31 | 46 | 0.002 | Chi square |
| MS Contin | 10 | 15 | 0.051 | Chi square |
| Median [Quartiles] dose of specific long-acting opiates (when single drug used only) | ||||
| Duragesic | 75 [50,87.5] | 50 [50,100] | 0.992 | Mann-Whitney U |
| Oxycontin | 20 [10,40] | 30 [10,40] | 0.485 | Mann-Whitney U |
| MS Contin | 30 [20,30] | 30 [20,60] | 0.838 | Mann-Whitney U |
| Patients with any use of long-acting opiates vs. ethnicity | ||||
| Ethnicity | Yes | No | ||
|
|
|
|
||
| African American | 84 | 87 | ||
| Hispanic | 45 | 75 | ||
| Caucasian | 20 | 25 | ||
| Asians | 3 | 15 | ||
| Not specified | 1 | 5 | ||
|
|
|
|||
| Total | 153 | 207 | 0.027 | Chi square |
| Patients with any use of long-acting opiates vs. gender | ||||
| Gender | Yes | No | ||
|
|
|
|
||
| Male | 86 | 108 | ||
| Female | 67 | 99 | ||
|
|
|
|||
| Total | 153 | 207 | 0.448 | Chi square |
| Patients adhering to pain regimen | ||||
| Visit 1 | 159 | 163 | ||
| Visit 2 | 163 | 160 | ||
| Visit 3 | 159 | 155 | 0.94 | Chi square |
| Percent of patients using any long-acting opiates (of ones reporting stage) | ||||
| Stage | All patients | Self Pay/Charity Care | Medicaid | |
|
|
|
|
|
|
| I | 12.5 | 0 | 25.0 | |
| II | 34.8 | 20.0 | 62.5 | |
| III | 32.7 | 30.0 | 36.4 | |
| IV | 46.7 | 41.8 | 51.5 | |
| p = 0.006 | p = 0.000 | p = 0.014 | Chi square | |
Multiple logistic regression of use of long acting opiates on both benefits and ethnicity showed benefits to be the only statistically significant predictor (p = 0.006) (Table 6). Even while controlling for ethnicity, the odds of prescribing long acting opiates for Medicaid patients was 2.4 times greater than the odds for CC/SP patients (p = 0.000). These data support our hypothesis that long acting opiate use in managing cancer pain in indigent patients is more frequent in patients who have Medicaid prescription coverage than Charity Care or Self Pay patients who have no prescription coverage.
Table 6. Predicting long-acting opiate use.
| Variable | Odds Ratio | p-value | test |
|---|---|---|---|
|
|
|
|
|
| Benefits | 2.716 | 0.000 | Logistic Regression |
| Accounting for race in the prediction (Race reference value = African-American) | |||
| Benefits | 2.443 | 0.006 | |
| Race: Asian or Hispanic | 0.606 | 0.158 | |
| Race: Caucasian | 1.007 | 0.987 | |
| Interaction: Benefits-Asian/Hispanic | 1.258 | 0.642 | |
| Interaction: Benefits-Caucasian | 0.889 | 0.886 | Logistic Regression |
Discussion
The data presented in this study support our hypothesis that indigent cancer patients treated in an inner city tertiary care medical center by Medical Oncologists have their pain treated differently depending on whether they have prescription coverage or not. Patients who have Medicaid prescription coverage received pain treatment with long acting opiates at a far greater rate than did indigent patients who had no insurance or prescription coverage. There were some differences in the two groups that merit discussion, however. The Charity Care/Self Pay group had a disproportionately higher percentage of Hispanic and Asian patients, while the Medicaid group had a significantly higher African American patient population. The distribution differences were likely due to United States residential status of the patients and their eligibility for Medicaid coverage. To provide support that this unbalance did not represent a confounding factor, we analyzed reported pain levels among the patients in the study and did not find statistically significant differences attributed to ethnicity. These findings are in agreement with reported pain assessment comparisons by patients and physicians, which found no differences between Hispanic and non-Hispanic Caucasian patients, even after controlling for multiple potential confounders (22). A study of physicians' attitudes towards cancer pain provides further support for lack of effect of ethnicity (30). In that study, similar rates of Hispanic patients (28%) as African-American patients (31%) received analgesics of insufficient strength to manage their pain (30). The data from our study showed that when treated individually in logistic regression analysis, both benefits and race had a statistically significant association with use of long acting opiates (Table 6). However, logistic regression analysis demonstrated that the addition of race produced only a negligible improvement in the model of long acting opiates as a function of benefits alone. The benefits category remains the dominant predictor of use of long acting opiates even when controlling for race and interaction of benefits and race (Table 5).
Pain distributions in our two payor groups were statistically indistinguishable (Table 3). This fact is a likely reflection on the payor category-independent attention to pain by the Medical Oncologist care providers in the practice, which ensured optimum pain control regardless of access to long-acting opiates. Nevertheless, the Medicaid group visited the Emergency Department at a higher rate and many of those visits included pain as a reason. However, the frequency of the most common reasons for emergency room visits was not statistically different in the two payor groups (Table 4). It is unclear whether the reasons for increased visits were financial, the influence of ethnic factors, or other unknown factors. The higher rate of substance use among the Medicaid patients in our study may suggest an additional potential reason for increased emergency department visits for pain. Others have also found that pain is among the top chief complaints of cancer patients visiting the ED and that 63% of visits result in admission to the hospital (31). However, other medical issues are common and also result in hospitalizations (31).
The rate of long term analgesic use among all patients increased with progressive stage, corresponding to increased stage-associated pain levels. This suggests that the first inclination to prescribe affordable short acting pain medications to uninsured patients was eventually trumped by an inflexible necessity to manage greater pain levels in later stages with long acting opiates, despite the hardship of out of pocket costs to the uninsured patient (Table 5).
In conclusion, our data support our hypothesis that the use of standard of care long acting opiates in indigent patients with cancer pain is significantly influenced by the availability of a prescription plan to the patient.
Supplementary Material
Acknowledgments
Supported by NJCCR 08-1096-CR-EO and NIH 1U10CA128506-01A1
Footnotes
Conflict of Interest statement: The investigators have no conflicts of interest.
This study has not been presented in a public forum.
References
- 1.Cleeland CS, Gonin R, Hatfield AK, et al. Pain and its treatment in outpatients with metastatic cancer. New England J Med. 1994;330:592–596. doi: 10.1056/NEJM199403033300902. [DOI] [PubMed] [Google Scholar]
- 2.Cleeland CS, Gonin R, Baez L, et al. Pain and treatment of pain in minority patients with cancer. The Eastern Cooperative Oncology Group Minority Outpatient Pain Study. Annals Internal Medicine. 1997;127:813–816. doi: 10.7326/0003-4819-127-9-199711010-00006. [DOI] [PubMed] [Google Scholar]
- 3.Teunissen SC, Wesker W, Kruitwagen C, et al. Symptom prevalence in patients with incurable cancer: a systematic review. J Pain and Symptom Management. 2007;34:94–104. doi: 10.1016/j.jpainsymman.2006.10.015. [DOI] [PubMed] [Google Scholar]
- 4.Dy SM. Evidence-Based Approaches to Pain in Advanced Cancer. The Cancer Journal. 2010;16:500–506. doi: 10.1097/PPO.0b013e3181f45853. [DOI] [PubMed] [Google Scholar]
- 5.Dy SM, Asch SM, Naeim A, et al. Evidence-based standards for cancer pain management. Journal of Clinical Oncology. 2008;26:3879–3885. doi: 10.1200/JCO.2007.15.9517. [DOI] [PubMed] [Google Scholar]
- 6.Chang VT, Hwang SS, Kasimis B. Longitudinal documentation of cancer pain management outcomes: a pilot study at a VA medical center. J Pain Symptom Management. 2002;24:494–505. doi: 10.1016/s0885-3924(02)00516-x. [DOI] [PubMed] [Google Scholar]
- 7.Dionne RA, Bartoshuk L, Mogil J, et al. Individual responder analyses for pain: does one pain scale fit all? Trends in Pharmacological Sciences. 2005;26:125–130. doi: 10.1016/j.tips.2005.01.009. [DOI] [PubMed] [Google Scholar]
- 8.Farrar JT, Polomano RC, Berlin JA, et al. A comparison of change in the 0-10 numeric rating scale to a pain relief scale and global medication performance scale in a short-term clinical trial of breakthrough pain intensity. Anesthesiology. 2010;112:1464–1472. doi: 10.1097/ALN.0b013e3181de0e6d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology for Adult Cancer Pain V I. 2007 http://www.nccn.org/professionals/physician_gls/PDF/pain.pdf.
- 10.Rhodes DJ, Koshy RC, Waterfield WC, et al. Feasibility of quantitative pain assessment in outpatient oncology practice. J Clin Oncol. 2001;19:501–508. doi: 10.1200/JCO.2001.19.2.501. [DOI] [PubMed] [Google Scholar]
- 11.Carr DB, Goudas LC, Balk EM, et al. Evidence report on the treatment of pain in cancer patients. J Natl Cancer Institute Monographs. 2004;32:23–31. doi: 10.1093/jncimonographs/lgh012. [DOI] [PubMed] [Google Scholar]
- 12.Du Pen SL, Du Pen AR, Polissar N, et al. Implementing guidelines for cancer pain management: results of a randomized controlled clinical trial. J Clin Oncol. 1999;17:361–370. doi: 10.1200/JCO.1999.17.1.361. [DOI] [PubMed] [Google Scholar]
- 13.Cleeland CS, Portenoy RK, Rue M, et al. Does an oral analgesic protocol improve pain control for patients with cancer? An intergroup study coordinated by the Eastern Cooperative Oncology Group. Annals of Oncology. 2005;16:972–980. doi: 10.1093/annonc/mdi191. [DOI] [PubMed] [Google Scholar]
- 14.Klepstad P, Kaasa S, Jystad A, et al. Immediate- or sustained-release morphine for dose finding during start of morphine to cancer patients: a randomized, double-blind trial. Pain. 2003;101:193–198. doi: 10.1016/s0304-3959(02)00328-7. [DOI] [PubMed] [Google Scholar]
- 15.Gatti A, Reale C, Occhioni R, et al. Standard therapy with opioids in chronic pain management: ORTIBER (ORamorph in TIBER) study. Clinical Drug Investigation. 2009;29(Suppl 1):17–23. doi: 10.2165/0044011-200929001-00003. [DOI] [PubMed] [Google Scholar]
- 16.Tassinari D, Sartori S, Tamburini E, et al. Transdermal fentanyl as a front-line approach to moderate-severe pain: a meta-analysis of randomized clinical trials. Journal of Palliative Care. 2009;25:172–180. [PubMed] [Google Scholar]
- 17.Rauck RL. What is the case for prescribing long-acting opioids over short-acting opioids for patients with chronic pain? A critical review. Pain Practice. 2009;9:468–479. doi: 10.1111/j.1533-2500.2009.00320.x. [DOI] [PubMed] [Google Scholar]
- 18.Bernabei R, Gambassi G, Lapane K, et al. Management of pain in elderly patients with cancer. SAGE Study Group. Systematic Assessment of Geriatric Drug Use via Epidemiology. JAMA. 1998;279:1877–1882. doi: 10.1001/jama.279.23.1877. [DOI] [PubMed] [Google Scholar]
- 19.Edwards RR, Fillingim RB, Ness TJ. Age-related differences in endogenous pain modulation: a comparison of diffuse noxious inhibitory controls in healthy older and younger adults. Pain. 2003;101:155–165. doi: 10.1016/s0304-3959(02)00324-x. [DOI] [PubMed] [Google Scholar]
- 20.Johnson VM, Teno JM, Bourbonniere M, et al. Palliative care needs of cancer patients in U.S. nursing homes. Journal of Palliative Medicine. 2005;8:273–279. doi: 10.1089/jpm.2005.8.273. [DOI] [PubMed] [Google Scholar]
- 21.Bartfield JM, Salluzzo RF, Raccio-Robak N, et al. Physician and patient factors influencing the treatment of low back pain. Pain. 1997;73:209–211. doi: 10.1016/S0304-3959(97)00107-3. [DOI] [PubMed] [Google Scholar]
- 22.Todd KH, Lee T, Hoffman JR. The effect of ethnicity on physician estimates of pain severity in patients with isolated extremity trauma. JAMA. 1994;271:925–928. [PubMed] [Google Scholar]
- 23.Cleeland CS, Mendoza TR, Wang XS, et al. Levels of symptom burden during chemotherapy for advanced lung cancer: differences between public hospitals and a tertiary cancer center. Journal of Clinical Oncology. 2011;29:2859–2865. doi: 10.1200/JCO.2010.33.4425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Green CR, Anderson KO, Baker TA, et al. The unequal burden of pain: confronting racial and ethnic disparities in pain. Pain Med. 2003;4:277–294. doi: 10.1046/j.1526-4637.2003.03034.x. [DOI] [PubMed] [Google Scholar]
- 25.Bryan M, De La Rosa N, Hill AM, et al. Influence of prescription benefits on pain control in patients with cancer. Pain Medicine. 2008;9:1148–1157. doi: 10.1111/j.1526-4637.2008.00427.x. [DOI] [PubMed] [Google Scholar]
- 26.Demidenko E. Sample size determination for logistic regression revisited. Statistics in Medicine. 2007;26:3385–3397. doi: 10.1002/sim.2771. [DOI] [PubMed] [Google Scholar]
- 27.Hosmer D, Lemeshow S. Applied Logistic Regression. New York, NY: John Wiley & Sons; 2000. pp. 339–347. [Google Scholar]
- 28.Fujimura T, Takahashi S, Kume H, et al. Cancer-related pain and quality of life in prostate cancer patients: assessment using the Functional Assessment of Prostate Cancer Therapy. International Journal of Urology. 2009;16:522–525. doi: 10.1111/j.1442-2042.2009.02291.x. [DOI] [PubMed] [Google Scholar]
- 29.Reyes-Gibby CC, Shete S, Yennurajalingam S, et al. Genetic and nongenetic covariates of pain severity in patients with adenocarcinoma of the pancreas: assessing the influence of cytokine genes. Journal of Pain & Symptom Management. 2009;38:894–902. doi: 10.1016/j.jpainsymman.2009.04.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Anderson KO, Mendoza TR, Valero V, et al. Minority cancer patients and their providers: pain management attitudes and practice. Cancer. 2000;88:1929–1938. [PubMed] [Google Scholar]
- 31.Mayer DK, Travers D, Wyss A, et al. Why do patients with cancer visit emergency departments? Results of a 2008 population study in North Carolina. Journal of Clinical Oncology. 2011;29:2683–2688. doi: 10.1200/JCO.2010.34.2816. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.