Description and Predictors of Direct and Indirect Costs of Pain Reported by Cancer Patients

Introduction 

The frequent nature of pain in cancer patients is well documented,1 and the deleterious effects of pain on patients' lives are difficult to exaggerate. Pain has been described as the most important2, 3 and most feared4 symptom cancer patients experience and has been linked to an assortment of decrements in quality of life, including psychological distress, diminished spiritual well-being, disturbed sleep, reduced appetite, and impairments in social functioning.4, 5, 6, 7, 8, 9 Although less well described, pain also exacts a serious economic toll on cancer patients, their families, and society at large.10, 11, 12, 13

The current national economic burden of cancer is estimated by the American Cancer Society at 107 billion dollars per year.14 While important information concerning the gross costs of cancer is readily available, little is known about the portion of cost resulting specifically from pain. Enumerating pain-related costs as a subset of these total cancer related costs is a complex task as cost categories are arbitrary and detailed financial data is frequently elusive.15 For the purposes of this study, we will distinguish between pain-related direct and indirect costs.16 The present study was a preliminary attempt to investigate the portion of direct and indirect costs attributable to pain and to investigate potential predictors of these costs.

Direct pain-related costs are those expenditures that are incurred from the direct treatment of pain, including analgesic medication, medical procedures and technology, hospitalizations, use of emergency department services, and physician office visits for pain.17 The American Cancer Society (ACS) estimates that the current economic burden due to direct medical expenditures for cancer treatment is US$ 37 billion.14 On an individual patient level, Hillner and colleagues18 estimated in 1988 that the average two-year direct medical expense incurred for the treatment of non-small-cell lung cancer was $48,000, including insurance payments, co-payments, and deductibles for 2 years after diagnosis or until death. Information about direct cost of pain specifically in cancer patients is less available.

Information has been recently published regarding the average wholesale costs per month to pharmacists in the treatment of a single patient for a range of analgesic medications.19 These costs ranged from $204 per month for transdermal fentanyl (Duragesic; 50 μg/hr) to $726 per month for subcutaneous hydromorphone (1 mg/ml @ 0.9 mg/hr). Oral and rectal morphine sulfate (MS Contin; 120 mg twice daily) cost $425 per month, and intravenous morphine (1 mg/ml @ 3.3 mg/hr) cost $381 per month. It should be noted that the doses of medications evaluated in the study were not equianalgesic doses.

In addition to the cost of analgesic medication, there are costs associated with its preparation and delivery. Labor necessary to prepare and administer certain analgesic regimens can substantially increase the total cost of an analgesic strategy. For example, the time required to prepare and administer morphine sulfate solution has been estimated to be twice that required for extended-release morphine sulfate tablets, making the total cost of the use of morphine sulfate solution substantially greater.20 Use of intravenous and subcutaneous infusions result in additional costs due to the equipment and disposables necessary for drug delivery. The 1994 cost of a patient receiving morphine via patient-controlled analgesia (PCA), including pump and disposables, was estimated to be $4,000 per month.21 The total cost of initial placement of an epidural catheter was reported in 1991 to range between $10,000 and $12,000,22, 23 and the 3-month cost of two commonly used implantable opioid delivery systems ranged between $15,600 to $16,300.23

A large portion of direct costs due to pain presumably results from hospitalizations necessitated by uncontrolled pain. One survey reported that 5% of inpatient cancer patients said they had previously been hospitalized because of pain.24 In a study of unscheduled admissions for uncontrolled pain at the City of Hope National Medical Center, pain was the second most frequent reason for unscheduled admission.25 The average length of stay for these hospitalizations was 12 days, and the average cost per hospital day was estimated at $1,800, resulting in a yearly cost of $5 million to the institution. Similarly, pain was responsible for 14% of unscheduled admissions to the MD Anderson Cancer Center and resulted in an average length of stay of 10.5 days at a cost of $1,200 per day, resulting in a yearly cost of approximately $5 million per year.10 Ultimately, all of society bears the costs of these services through taxes and insurance premiums, but pain patients and their families assume a portion of these direct medical costs through deductibles, co-payments, and services not covered by third parties.26 Although these studies give some indication that pain results in substantial direct medical expense, there is no information available about direct medical costs associated with pain-related emergency department visits or physician office visits resulting from pain, and likewise little has been done to enumerate the indirect costs of pain.

Indirect pain-related costs are all other costs resulting as a consequence of pain, including, for example, money spent on transportation or extra household help.17 Indirect costs include expenditures for numerous items and services outside direct medical care and may refer to loss of earnings or productivity in addition to expenditures incurred as a result of pain. No published study has addressed indirect costs related specifically to pain, but two sources give reason to believe that indirect costs resulting from cancer generally are substantial. The ACS estimates that the current indirect cost of cancer is $70 billion. Stommel et al. estimated out-of-pocket direct and indirect costs in 192 cancer patients, finding the average cost was approximately $5,400 for a 3-month period, 79% of which were indirect costs resulting from lost wages.26 Indirect costs may be neglected in research because costs due to non-drug treatments for pain and other complementary modalities of pain control such as psychoanalysis, support groups, books, videotapes, and meditation are less likely to be covered by insurance and are received through multiple sources without the direct knowledge or documentation of physicians.

The purpose of this study was to describe direct and indirect costs associated with cancer pain and to examine potential predictors of these costs. The study surveyed cancer outpatients about the occurrence of pain-related hospitalizations, emergency department visits, physician office visits, and use of analgesic medications, which comprise direct costs, and about money spent on pain-related transportation, complementary methods to improve pain, educational materials, over-the counter medication, domestic support, and childcare, which comprise indirect costs. Furthermore, the study examined age, marital status, race, income level, pain severity, and pain interference as predictors of direct and indirect costs.

Methods 

Results 

Sample 

A total of 373 cancer outpatients were sampled. Of those, 144 cancer patients (39%) reported experiencing cancer-related pain and completed the study questionnaires. Table 1 shows the sample characteristics of those patients who reported experiencing pain. The mean age of participants was 54 years. Most participants (74%) were women, and the majority of participants (85%) were white with at least some college education (75%). The most common cancer diagnosis was breast cancer (40%), followed by gastrointestinal cancer (16%) and genitourinary cancer (15%). Patient scores for the severity of worst pain experienced in the past 24 hours ranged from 1 to 10 with a mean of 4.4 (SD 2.5), which is considered “moderate” pain when using a scale of 0 (no pain) through 10 (pain as bad as you can imagine).35 Thirty-three of the patients (23%) experienced breakthrough pain according to criteria set by the researchers.

Table 1. Patient Characteristics (n = 144)

	Age	M = 53.84 (SD = 13.23)
	Education (≥ Some college)	74.6%
	Sex (women)	74.3%
	Household Income
	Under $14,999	11.7%
	$15,000–49,999	38.7%
	$50,000–74,999	22.6%
	$75,000–99,999	6.6%
	$100,000–149,999	10.2%
	Over $150,000	10.2%
	Marital Status (Married)	62.5%
	Race
	Caucasian	85.4%
	African American (Non-Hispanic)	6.3%
	Asian American	2.8%
	Hispanic	2.1%
	Native American	2.1%
	Cancer Type
	Breast	40.3%
	Gastrointestinal	16.0%
	Genitourinary	15.3%
	Lymphoma	6.9%
	Lung	5.6%
	Leukemia	3.5%
	Unknown Primary	3.5%
	Sarcoma	3.5%
	Multiple Myeloma	2.8%
	Head and Neck	2.1%
	Melanoma	0.7%
	Pain Severity in the Past 24 Hours
	Mild Pain	52.8%
	Moderate Pain	23.6%
	Severe Pain	22.2%
	Patients With Breakthrough Pain	23.0%
	Type of Analgesics Patients Used According to World Health Organization (WHO) Analgesic Ladder
	Step 1 or Less (acetaminophen, aspirin, NSAIDs)	47.2%
	Step 2 (codeine, hydrocodone, oxycodone, propoxyphene)	29.9%
	Step 3 (levorphanol, methadone, fentanyl, hydromorphone, morphine)	22.9%

Total Pain-Related Costs 

Table 2 shows the 3-month incidence of costs experienced, the maximum monthly amount reported by an individual patient, and the average monthly amount for both direct and indirect pain-related costs across all patients. Seventy-six percent (76%) of the patients had experienced at least one pain-related cost, whether it was direct or indirect, resulting in an average monthly cost of $891/month per patient. Ninety-three percent (93%) of the total monthly cost estimate was attributed to direct costs, and 71% of the total costs were attributed specifically to the cost of pain-related hospitalizations.

Table 2. Direct and Indirect Pain-Related Costs Reported by Cancer Patients (n = 144)

		3-Month Incidence	Maximum ($/Month)	Mean (SD) ($/Month/Patient)
	Direct Costs
	Analgesic Medication	55.2%	3,575.10	81.76 (317.40)
	Medical Visits	44.6%	1,100.00	89.58 (161.17)
	Emergency Department Visits	12.7%	400.00	23.15 (65.64)
	Hospitalizations	7.9%	17,050.00	631.48 (2,339.06)
	Total Direct Costs	69.1%	18,702.80	825.97 (2,500.36)
	Indirect Costs
	Transportation	41.2%	273.00	10.66 (32.97)
	Over-the-Counter Medications	38.8%	133.22	7.54 (17.11)
	Complementary Medicine	20.0%	233.33	10.78 (35.11)
	Educational Materials	12.1%	100.00	3.27 (35.11)
	Counseling or Psychotherapy	13.0%	233.33	4.35 (26.12)
	Household	9.1%	1,333.33	24.70 (127.50)
	Childcare	0.6%	104.17	0.72 (8.68)
	Total Indirect Costs	57.0%	1,340.00	61.09 (149.03)
	Total Pain-Related Costs	75.8%	18,753.80	891.10 (2,500.36)a

a The sum of the total direct costs and total indirect costs vary slightly from the total pain-related costs because data were missing from one patient for a subcategory in the indirect costs. Total costs were calculated on patients who had complete data for both categories.

Discussion 

In our study, cancer outpatients with pain experienced an average of about $900 per month for total pain-related costs, or approximately $10,000 per year. Interestingly, 93% of pain-related cost was associated with direct medical expenses, with hospitalizations overwhelmingly being the greatest direct medical expense, even though they were the least frequently occurring. These numbers are not surprising given the disproportionately high cost of inpatient medical care for cancer patients experiencing uncontrolled pain.10, 25 In our study, direct pain-related medical expenditures comprised 93% of total pain-related costs, a result that differed considerably from the ACS report that direct costs comprise 35% of total costs. However, in contrast to the ACS, we did not collect data regarding indirect costs associated with lost productivity. Moreover, the costs reported by the ACS were not limited to pain-related costs.

The pain-related costs reported in our study underestimate the true economic costs of cancer pain. Our study addressed an outpatient population, which is likely to disproportionately represent healthier cancer patients and therefore, those less likely to have pain. Those too ill to easily come for outpatient visits, those followed by hospice programs, and those hospitalized at the time of the survey could not be studied. Furthermore, a measure of lost patient/caregiver productivity was not included in our survey, and medication costs, when dosing was not known, were conservatively estimated.

Prescription medication was the most frequently occurring direct pain-related medical expense, followed closely by the cost of medical visits. Emergency department visits and hospitalizations for pain occurred relatively infrequently. Because patients and the health care system share these direct pain-related costs, they represent a significant financial burden to society. Moreover, cancer patients incur a sizeable financial burden due to deductibles, co-payments, and charges for services not covered by third party reimbursement. For those patients without health insurance, the complete burden of care falls to them and to the institutions that serve them.

There were a number of factors that were found to be significantly related to, and thus candidate predictors of, pain-related medical costs. Although the r-values for the predictors of pain-related costs were relatively weak, this may have been a result of the small sample size. Despite the small effect size, however, these findings are worth noting due to the large impact of pain-related costs on a grand scale. Patients who had greater pain, whose pain interfered with their ability to function, who had breakthrough pain, who had less income, and who were relatively younger incurred more direct pain-related expenses. The fact that pain intensity, pain interference, and presence of breakthrough pain predicts greater direct pain-related expense is expected, but the relationship of income level and age is not necessarily straightforward and deserves careful consideration and further evaluation. Variations in pain management practice that occur in relation to the income level of patients need to be examined further since people with lower income may be less able to pay for the common treatments for pain, making them potentially vulnerable to pain crises that require hospitalization and other relatively more expensive scenarios. The relationship of age to direct pain-related expenses is equally intriguing, and several possibilities need to be tested to explain the relationship. It may be that older patients are treated less aggressively or it may be that older cancer patients respond to pain in ways that reduce the likelihood of hospitalization or the utilization of other expensive analgesic procedures.

Indirect pain-related costs targeted by our study included those resulting from transportation, over-the-counter medication, household help, childcare, educational materials, and complementary medicine. Cancer patients in our study reported spending an average of $53 per month on these items altogether, or approximately $640 per year. This figure is magnified when one understands that indirect costs are typically borne entirely by the patient or, in other words, are out-of-pocket expenses. Assuming a median household income of $37,000 per year,36 the indirect costs described by our study account for approximately 2% of the average annual household income. Furthermore, it should be noted that indirect expenses are in addition to out-of-pocket expenses resulting from direct medical care, which alone may be substantial. Moore studied a convenience sample of 20 adult cancer patients receiving chemotherapy and estimated that the out-of-pocket expenses for these patients ranged up to $3,000 during a course of chemotherapy.17 Even seemingly minor indirect costs in the context of other out-of-pocket expenses and potential losses of income from reduced occupational status can be serious.

The most frequent indirect cost involved transportation, followed closely by costs resulting from over-the-counter medication. Interestingly enough, the third most frequent indirect cost resulted from complementary medical treatments of pain, reported by 20% of participants, which is consistent with recent reports of the frequent use of complementary medical treatments by cancer patients.37, 38, 39, 40 Costs resulting from counseling or psychotherapy, educational materials, and household help occurred in 9–10% of patients, while childcare resulted in an indirect expenditure in less than 1% of patients. As noted, there was a fairly high need for assistance with household chores, but of the patients who received help with household chores, only a small portion actually paid for this help. This supports the idea that functional ability is impaired in patients suffering from cancer pain but highlights a limitation of the present study in that we did not consider indirect costs resulting from unpaid labor and lost time at work by patients, friends, and relatives. A previous report show that when dollar figures are applied to hours of help provided to cancer patients by family and friends these costs become the largest single cost.26

We believed marital status and income level would predict the amount of indirect expenses people reported. However, our data did not support this hypothesis. Pain intensity, pain interference, and presence of breakthrough pain predicted higher indirect expenses, suggesting that as pain intrudes on the daily lives of cancer patients they begin to incur expenses for issues outside of the direct medical treatment of pain, such as seeking complementary treatments, educational materials, and over-the-counter medication.

Given the preliminary picture of pain-related costs provided by this study, it is reasonable to ask if these expenses can be minimized. Cost shifting, which occurs when costs to the health care provider are shifted to the patient in the form of outpatient or home care, has been identified as an important consideration of pain management that may help reduce the cost of pain care.21 Although these shifts in care may appear to reduce costs in relation to the institution or direct costs from which the cost was “shifted”, it may also reduce the quality of care to patients and increase indirect and out-of-pocket costs to patients and their families. Whether indirect and direct costs are considered in relation to quality of care can make a major difference to the final cost/benefit analysis of a particular treatment.

Another possibility involves the reduction of unnecessary hospitalizations for pain through better pain management and patient education. There is good preliminary evidence that direct pain-related costs due to unnecessary hospitalizations can be minimized through improvement in pain management practices. Preventable readmissions, which occur typically within one month of discharge, may be targeted through patient education and discharge planning.41 Grant et al. reported that 26% of patients experiencing unscheduled hospitalizations for pain had a previous hospital admission for pain, 54% of which occurred within 2 weeks of discharge.25 Following a pain initiative that included implementation of pain documentation procedures and education of staff and patients, the estimated yearly cost of unscheduled admissions for pain was reduced by approximately $1 million dollars.

Another possible way to reduce pain-related direct medical costs involves the expeditious use of expensive technology in gaining control of cancer pain. Jennett described a framework for determining if the use of expensive analgesic technology is unnecessary, unsuccessful, unsafe, unkind, or unwise.42 Using this framework, Ferrell and Rhiner found that 52% of patients were advanced to parenteral morphine before the less expensive option of oral morphine was tried, and among those patients who had tried oral morphine before advancing to the parenteral route, 52% had received only mild or moderate analgesics.13 Fourteen percent of patients receiving parenteral morphine had been changed directly from acetaminophen and codeine. Considering less expensive oral analgesics before proceeding to other expensive analgesic strategies may result in considerable costs savings while maintaining an acceptable level of pain control.

In summary, the present study provides a preliminary picture of the direct and indirect costs related to pain in cancer patients and highlights predictors of these expenses. To interpret these findings, it is important to understand the limitations of the study. One limitation of our study was that the sample was small and nonrandom. Generalization of study findings should consider the fact that the survey took place in four outpatient clinics with a heterogeneous set of cancer patients. Moreover, the study was not adequately powered to determine whether demographic or socioeconomic characteristics predict pain-related costs. Therefore, our findings that these characteristics were not predictive of pain-related costs should be interpreted with caution. The extent to which direct and indirect pain-related costs are affected by socioeconomic and disease characteristics requires future investigation. Another limitation of our study is the reliance on cancer patients to retrospectively identify costs attributable to pain. It is not clear to what extent cancer outpatients are able to accurately recall medical events and the issues prompting particular medical events, but previous studies suggest that patients tend to underreport medical utilization,43, 44 suggesting our findings may be conservative. This is a difficult issue because patients may attribute, for example, a hospitalization to pain when it actually occurred because of a medical problem that would have required hospitalization regardless of pain experienced. On the other hand, it is also possible that patient attributions may identify medical procedures that would not have occurred or that would have been executed in less expensive forms, such as outpatient treatment versus inpatient care, had pain not been present or had it been minimal. Furthermore, it is difficult to imagine a feasible research design that identified many indirect costs without some degree of patient report. We look forward to the publication of studies that employ methods that give more exact cost estimations that are less dependent on patient report.