Abstract
Background
Childhood cancers typically require rigorous treatment at specialized centers in urban areas, which can create substantial challenges for families residing in remote communities. We evaluated the impact of residence and travel time on the burden of care for families of childhood cancer patients.
Procedure
We conducted a cross-sectional, self-administered survey of 354 caregivers of pediatric cancer patients at a children’s hospital serving a seven state area. Measures included the impact of cancer treatment on relocation, employment, schooling and finances. We evaluated these domains by rural/urban residence and travel time (>1 hour and >2 hours) to the hospital in multivariable regression models.
Results
Of the 29% of caregivers who reported moving residences since their child was diagnosed, 33% reported that the move was due to their child’s cancer. Rural and remote (e.g., >1 hour travel time) caregivers missed more days of work during the first month after diagnosis than did urban and local caregivers, however, these differences did not persist over the first six months of therapy. One-third of caregivers reported quitting or changing jobs as a direct result of their child being diagnosed with cancer. Rural respondents had greater out-of-pocket travel expenses and reported a significantly greater perceived financial burden. Rural patients missed more school days and were at an increased risk of having to repeat a grade.
Conclusions
Childhood cancer has an appreciable impact on the lives of patients and caregivers. The burden is greater for those living far from a treatment center.
Keywords: Pediatric Oncology, Disparities, Burden of Care, Rural
Introduction
Childhood cancers typically require treatment at specialized tertiary care centers that offer dedicated pediatric services and a multidisciplinary approach to care. Such specialized centers tend to be located in large urban areas and treat patients that come from diverse geographic areas and, often, great distances. Approximately one-fifth of Americans live in rural areas1 and many have travel times of greater than one hour to a designated National Cancer Institute Designated Cancer Center and other specialized treatment facilities2. Studies have shown that balancing a rigorous treatment schedule with school, work and family obligations can be financially taxing for families affected by pediatric cancer3–9.
The burden associated with childhood cancer care may be greater for families from rural communities and for those traveling great distances to primary cancer centers for therapy10. While there are several descriptive studies that assess the burden experienced by families during childhood cancer treatment6,11–15, little is known about the effects of place of residence or travel time to therapy on that burden in the United States. Further, a more comprehensive understanding of burdens beyond the direct economic effects from therapy, such as the impact on parental employment and patient education or from treatment-related factors such as emergency medical transport, is needed to better meet the supportive care needs for families. In the current study, we surveyed caretakers of pediatric cancer patients treated at a tertiary care center about the impact of cancer and its treatment on relocation, employment, schooling, and finances. Our analyses focused on determining whether residence (rural vs. urban) and travel time to the treatment center (>1 hour and >2 hours) imparts a greater care burden for these families.
Methods
Study Center
The study setting was the oncology clinic at Primary Children’s Hospital (PCH) in Salt Lake City. PCH serves as the primary pediatric tertiary care center for one of the largest areas in the continental United States served by a single children’s hospital.
Participant Sample
Primary Caretaker Survey Sample: All primary caretakers of patients, diagnosed with a cancer between 0–18 years of age, who were at least three months from diagnosis, who were less than 60 months from the end of therapy, and who received care in the oncology clinic from Jan 1, 2010 to Dec 31, 2012 were eligible for participation. Of 449 caregivers approached to participate during this timeframe, 356 completed surveys. Two surveys were excluded due to a lack of travel time and rurality data, leaving 354 evaluable surveys (308 English and 46 Spanish language) for a response rate of 79%. To assure sample representativeness, we compared participant demographics to those for all outpatient visits using medical records for the same two-year period.
Survey Methods
Institutional Review Board approval was obtained. The burden of care, from the perspective of the patients’ primary caregiver, was evaluated using a 48-item self-administered survey. Outcomes of interest included financial burden (travel costs, relocations, missed work, missed school and perception of financial burden) and treatment-related factors, including utilization of medical transport. The survey was available in English and Spanish.
Both rurality and travel time were used as independent variables to allow for differential assessment of the impact of having easier access to the primary treatment center as well as having access to other resources more commonly found in urban areas. Rural vs. urban residence at diagnosis was classified by Rural-Urban Commuting Area (RUCA) Codes16,17, a ZIP code-based geographic taxonomy utilizing the Bureau of Census urban area and urban cluster definitions and work commuting flows to characterize areas regarding their rural and urban status.
Travel times were calculated using MapQuest. Travel time of greater/equal to or less than one hour from respondents’ residence to PCH was categorized because this time cutoff is used to direct patients with neutropenic fever to local hospitals, rather than PCH, for urgent care. Travel time of greater/equal to or less than two hours to PCH was categorized because this travel time encompasses the major metropolitan areas in the immediate region of the primary treatment center.
Cancer diagnoses were grouped into acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), brain tumors, and solid tumors, creating groups with similar treatment schedules.
Statistical Analysis
For comparison of the geographic groups on patient demographics, two-sample t-tests for continuous variables, Wilcoxon-Mann-Whitney tests for ordered categorical variables, and chi-square or Fisher’s exact test, as appropriate, for unordered categorical variables were used. For dichotomous study outcomes, a multivariable Poisson regression was used, using time since diagnosis as the time variable in rate estimates. For continuous study outcomes, a multivariable linear regression was used adjusting for time since diagnosis. To compare groups for the ordered categorical variable “number of work days missed for each clinic visit”, a multivariable logistic regression model was fitted using the rural/urban and travel time groups as the outcome variables and ‘days missed’ categories as a continuous predictor variable, thus utilizing a linear trend test, as advocated by Woodward18.
Multivariable models included all variables with p-value <0.20 for association with urban/rural residence or travel time from univariate analysis (Table I). For all multivariable models, there were sufficient sample sizes and sufficient number of events to avoid over-fitting. The final models were adjusted for age at diagnosis, diagnosis, language, race, ethnicity, education and religion, and incorporated time since diagnosis. Missing data was not imputed. Rather, list-wise deletion was utilized (subjects with missing data were excluded from analyses for which the missing data were part of the model). All reported p-values are two sided. Statistical analyses were performed using Stata (College Station, TX).
Table I.
Characteristics of Childhood Cancer Patients and Caregivers by Urban/Rural Residence and by Travel Time to Treatment Center
| Overall [n=354] | Residencea | Travel time to treatment center | Travel time to treatment center | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Rural (18.6%) [n=66] | Urban (81.4%) [n=288] | p-valueb | ≥ 1hr (39.3%) [n=139] | <1hr (60.7%) [n=215] | p-valueb | ≥ 2hr (25.7%) [n=91] | <2hr (74.3%) [n=263] | p-valueb | ||
| Age of cancer patient, yrs, mean ± sd | 8.8±5.6 | 9.6±6.2 | 8.7±5.5 | 0.22 | 9.2±6.0 | 8.6±5.4 | 0.34 | 9.4±6.1 | 8.6±5.5 | 0.27 |
| Time from dx to survey-yrs, mean ± sd | 2.2±2.2 | 2.0±1.8 | 2.2±2.3 | 0.48 | 2.2±2.3 | 2.1±2.2 | 0.86 | 2.0±1.9 | 2.2±2.3 | 0.49 |
| Male (patient) n(%) | 195(56) | 36(55) | 159(57) | 0.86 | 79(57) | 116(56) | 0.79 | 50(56) | 145(57) | 0.86 |
| Travel Time, min, mean ± sd | 110±167 | 212±104 | 87±169 | <0.001 | 225±220 | 35.8±13.9 | <0.001 | 300±241 | 44.5±23.7 | <0.01 |
| Hispanic ethnicity n(%) | 71(20) | 11(17) | 60(21) | 0.43 | 22(16) | 49(23) | 0.09b | 52(20) | 19(21) | 0.86 |
| White race n(%) | 292(82) | 59(89) | 233(81) | 0.10b | 121(87) | 171(80) | 0.07 | 76(84) | 216(82) | 0.76 |
| Cancer Type, (n%) | 0.05b | 0.06b | 0.002b | |||||||
| Acute Lymphoblastic Leukemia (ALL) | 157(44) | 30(45) | 127(44) | 58(42) | 99(46) | 35(38) | 122(46) | |||
| Acute Myeloid Leukemia (AML) | 18(5) | 3(5) | 15(5) | 6(4) | 12(6) | 3(3) | 15(6) | |||
| Brain Tumor | 45(13) | 2(3) | 43(15) | 12(9) | 33(16) | 5(5) | 40(15) | |||
| Solid Tumor | 132(37) | 31(47) | 101(35) | 63(45) | 69(32) | 48(52) | 84(32) | |||
| Primary Language (caregiver) | ||||||||||
| English n(%) | 308(87) | 62(94) | 246(85) | 0.07b | 128(92) | 180(84) | 0.02b | 81(89) | 227(86) | 0.51 |
| Spanish n(%) | 46(13) | 4(6) | 42(15) | 11(8) | 35(16) | 10(11) | 36(13) | |||
| Household Income Category n(%) | 0.21 | 0.79 | 0.82 | |||||||
| <$19,999 | 44(13) | 5(8) | 39(15) | 18(14) | 26(13) | 12(14) | 32(13) | |||
| $20,000 – $39,000 | 82(25) | 21(34) | 61(23) | 30(23) | 52(27) | 22(26) | 60(25) | |||
| $40,000 – $59,000 | 71(22) | 12(20) | 59(22) | 33(26) | 38(19) | 18(21) | 53(22) | |||
| $60,000 – $79,000 | 52(16) | 11(18) | 41(16) | 19(15) | 33(17) | 13(15) | 39(16) | |||
| $80,000 – $99,000 | 41(13) | 9(15) | 32(12) | 17(13) | 24(12) | 13(15) | 28(12) | |||
| Over $100,000 | 35(11) | 3(5) | 32(12) | 12(9) | 23(12) | 6(7) | 29(12) | |||
| Preferred not to answerc n | 29 | 5 | 24 | 0.90c | 10 | 19 | 0.70c | 7 | 22 | 0.85c |
| Highest Level of Education of Caregiver n(%) | 0.29 | 0.07b | 0.59 | |||||||
| Grades 1 through 8 | 23(7) | 2(3) | 21(7) | 4(3) | 19(9) | 4(5) | 19(7) | |||
| Grades 9 through 11 (some high school) | 16(5) | 3(5) | 13(5) | 5(4) | 11(5) | 5(6) | 11(4) | |||
| High School or GED | 43(12) | 10(16) | 33(12) | 23(17) | 20(10) | 13(15) | 30(12) | |||
| Some college, associate degree or vocational tech school | 116(34) | 26(42) | 90(32) | 45(34) | 71(34) | 32(37) | 84(33) | |||
| College graduate or more | 146(42) | 21(34) | 125(44) | 57(43) | 89(42) | 32(37) | 114(44) | |||
| Preferred not to answerc n | 10 | 4 | 6 | 0.09 c | 5 | 5 | 0.49c | 5 | 5 | 0.09 c |
| Insurance Status (patient) n(%)d | 0.68 | 0.60 | 0.71 | |||||||
| No Insurance | 6(2) | 1(2) | 5(2) | 0.90 | 2(1) | 4(2) | 0.76 | 2(2) | 4(2) | 0.67 |
| Medicaid Insured | 128(36) | 25(38) | 103(36) | 0.75 | 52(37) | 76(35) | 0.69 | 33(36) | 95(36) | 0.98 |
| Religion of Caregiver n(%) | 0.27 | 0.26 | 0.21 | |||||||
| Religion LDS | 218(63) | 33(51) | 185(66) | 0.03b | 83(61) | 135(64) | 0.49 | 50(56) | 168(65) | 0.13b |
| Religion Catholic | 53(15) | 13(20) | 40(14) | 19(14) | 34(16) | 15(17) | 38(15) | |||
| Religion Other | 41(12) | 11(17) | 30(11) | 22(16) | 19(9) | 16(18) | 25(10) | |||
| Not religious n(%) | 35(10) | 8(12) | 27(10) | 13(9) | 22(10) | 8(9) | 27(11) | |||
| Religious Service Attendance Frequency (caregiver) | 0.01b | 0.03b | 0.62 | |||||||
Rural/urban residence was based on Rural-Urban Computing Area (RUCA) codes at the zip code level. RUCA codes were developed by the University of Washington Rural Health Research Center and the USDA Economic Research Service (ERS), with the support of the federal Health Resource and Service Administration’s Office of Rural Health Policy and the ERS using standard Census Bureau urbanized area and urban cluster definitions in combination with work commuting data to characterize census tracts and later zip codes. The 10 RUCA categories were aggregated into urban (1–3) and rural (4–10) as recommended by the WWAMI (Washington, Wyoming, Alaska, Montana, and Idaho) Rural Health Research Center.
p-values reflect the significance of the differences between comparison groups with respect to each variable. p-Values for variables with >2 options (ie cancer type, income, education, religion) reflect the differences in distribution among the comparison groups of interest. Variables found to differ among comparison groups with a p-value of <0.2 were included in multivariable models
p-values reflect the significance of the differences between comparison groups with respect to number of subjects with missing data. Missing data includes the following: Age (n=6), Gender (n=8), Ethnicity (n=7), Diagnosis (n=2), Income (n=29), Education (n=10), Religion (n=7), Health Insurance (1).
All patients not listed as Medicaid insured or uninsured had private or military insurance
Results
Caregiver/ Patient Characteristics
Mean patient ages were 6.7 years (sd 5.4, median 5, IQR 2,11) at cancer diagnosis and 8.8 years (sd 5.6, median 7.5, IQR 4,13) at the time of the survey, with a mean time from diagnosis to survey of 2.2 years (sd 2.2, median 1.5, IQR 0.4,3.1). The most common cancer diagnosis was ALL (44%), followed by solid tumors (37%), brain tumors (13%) and AML (5%) (Table I). Respondents did not differ significantly from those who refused participation in terms of age, gender, ethnicity, cancer, or time from diagnosis. A review of all outpatient medical records over the same two-year period found that the surveyed cohort is generally representative of our patient population (36% with ALL diagnosis, 54% male, 13.1% Hispanic, a mean age of 9.7 years and 27% of patients traveling >2 hours to the primary treatment center),
Travel and Transportation
Rural dwellers accounted for 18.6% of the respondents (Table I). Mean travel time from patients’ homes to the treatment center was 110 (sd 167, median 52, IQR 29,131) minutes with 39% and 26% reporting >1 and >2 hours travel times, respectively. Over 98% of caregivers utilized a car as transportation for routine clinical visits and hospitalizations, while a small fraction used a taxi (0.8%), train (3.3%), bus (0.5%) or airplane (1.7%).
There was no difference in the proportion of rural versus urban respondents whose child required a transfer by ambulance at any time since their cancer diagnosis. However, travel of >1hr was found to increase the risk of requiring an ambulance (unadjusted percents 30% for >1 hour vs. 20% for <1hr, aRR 1.6, 95% CI 1.1 to 2.6, p=0.03). Likewise, a travel time of >2 hours increased the risk of ambulance utilization (unadjusted percents, 28% vs 23%, respectively), although this was marginally significant in adjusted models (aRR 1.6, 95% CI 1.0 to 2.6, p=0.06). Rurality was strongly associated with increased utilization of emergency air (airplane/helicopter) transport with 23% of rural respondents reported utilizing air transport in contrast to 5% of urban respondents (aRR 5.9, 95% CI 2.7 to 13.1, p<0.001). This difference was partially explained by the effect of increasing travel time as a travel time of >1hr was associated with an increased risk of emergency air transport (unadjusted percents: 15% for >1hr vs. 4% for <1hr, aRR 3.8, 95% CI 1.7 to 8.5, p=0.001). Likewise, 21% of those traveling >2hrs vs. 4% for <2hrs, utilized emergency air transport (aRR 7.2, 95% CI 3.5 to 14.5, p<0.001). (Table II).
Table II.
Multivariable Poisson Regression Models for Dichotomous Outcomes: Use of Emergency Transportation and Burden in Childhood Cancer Patients' Families by Urban/Rural Residence and Travel Time
| Outcome | Rural vs. Urban Residence | Travel time to treatment center ≥ 1hr vs. < 1hr | Travel time to treatment center ≥ 2hrs vs. < 2hrs | |||
|---|---|---|---|---|---|---|
|
| ||||||
| Rate Ratio | 95% CI | Rate Ratio | 95% CI | Rate Ratio | 95% CI | |
| Emergency Transportation Usea | ||||||
| Any ambulance | 1.6 | 0.9 to 2.8 | 1.6 | 1.1 to 2.6 | 1.6 | 1.0 to 2.5 |
| Any emergency air transport | 5.9 | 2.7 to 13.1 | 3.8 | 1.7 to 8.5 | 7.2 | 3.5 to14.5 |
| Burden Measuresa | ||||||
| Relocated residence due to cancer | 1.9 | 0.9 to 4.3 | 1.7 | 0.8 to 3.9 | 2.1 | 1.0 to 4.4 |
| Caregiver quit work | 1.0 | 0.6 to 1.7 | 0.9 | 0.6 to 1.3 | 0.9 | 0.6 to 1.5 |
| Patient Held back grade in school | 3.0 | 1.1 to 8.3 | 0.9 | 0.2 to 3.1 | 1.5 | 0.5 to 4.7 |
Each row represents different multivariable Poisson regression model adjusted for age at diagnosis, diagnosis, language, race, ethnicity, education and religion. Time since diagnosis was incorporated as the time variable in the Poisson models.
More rural patients (66%) received some type of healthcare outside of PCH at least one time since their diagnosis compared to urban patients (46%). Likewise, 70% of patients traveling >2 hours, compared to 43% for <2 hours, received care at another facility.
Relocation of Residence
Overall, 29% of caregivers reported moving residences since their child was diagnosed with cancer. Of these, 33% reported that the move was directly due to their child having cancer, citing ‘being closer to the treatment center’ and ‘being closer to family support’ as the most common reasons. Greater travel time increased the likelihood of the move being “directly due to cancer” (aRR 1.7, p=0.2 for travel time >1 hour and RR 2.1, p=0.05 for travel time >2 hours). (Table II)
Employment / Financial Burden
Caregivers missed a mean of 17.3 days (sd 32, median 12, IQR 7,20) of work during the first month after diagnosis, with rural families missing significantly more days than urban respondents (24 vs. 16 days, respectively, adjusted mean difference 10.1, 95% CI 0.66 to 20.0, p=0.04). While the number of missed workdays was not significantly different for >1 hour of travel, the difference reached significance for those traveling >2hrs (24 days for >2 hours vs. 15 days for <2 hours, adjusted mean difference 10.5, 95% CI 2.1 to 18.8, p=0.01). When assessing monthly missed days during the first six months (mean 14, sd 31, median 5, IQR 2,15), there were no significant differences either by rural/urban residence or by travel time.
The number of work days missed for each clinic appointment included choices for none, half, one, two or more than two days. There was a significant increase in workdays missed for each visit for rural respondents (adjusted p_trend = 0.02), those traveling >1hr (adjusted p_trend=0.002), and those traveling >2hrs (adjusted p_trend=0.001).
Over one-third of respondents (36%) reported that at least one caregiver quit or changed jobs as a direct result of their child being diagnosed with cancer. When this was examined by residence and travel time, no significant differences emerged.
Another contributor to a caregiver’s financial burden was the direct, out-of-pocket travel costs, including gasoline, airline or bus tickets, and hotel stays. The mean cost, per visit, was $60.21 (sd $100.6) for the entire cohort, but varied widely based on location of residence (rural = $138 vs. urban = $42, p<0.001, travel time >1hr = $124 vs. <1hr, travel time = $19, p<0.001, travel time >2hrs = $168 vs. <2hrs = $23, p<0.001). (adjusted mean differences shown in Table III)
Table III.
Multivariablea Linear Regression Models for Continuous Outcomes: Financial Burden and Missed Work and School for Childhood Cancer Patients' and Families, by Urban/Rural Residence and Travel Time
| Rural vs. Urban Residence | Travel time to treatment center ≥ 1hr vs. < 1hr | Travel time to treatment center ≥ 2hrs vs. < 2hrs | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| Outcome | Adjusted mean difference | 95% CI | Adjusted mean difference | 95% CI | Adjusted mean difference | 95% CI |
| Perception of Financial burden (Visual Analogue Scale, 0–100) | 9.6 | 0.3 to 18.8 | 3.3 | −4.1 to 10.8 | 8.9* | 0.7 to 17.1 |
| Travel expenses, $ | 96 | 71 to 121 | 104 | 86, to 123 | 144 | 125 to 163 |
| Missed work during first month of illness, days | 10.6 | 1.1 to 20.1 | 6.3 | −1.3 to 13.9 | 10.5 | 2.1 to 18.8 |
| Missed work during first 6 months of illness, days | −2.1 | −11.2 to 7.0 | −1.0 | −8.4 to 6.3 | 2.3 | −5.8 to 10.3 |
| Missed school during first 6 months of illness, days | 32.7 | 2.0 to 63.4 | 19.4 | −11.1 to 48.0 | 32.8 | 3.6 to 62 |
Each row represents a different multivariable linear regression model adjusted for age at diagnosis, diagnosis, language, race, ethnicity, education and religion, as well as time since diagnosis.
Participants were asked to rate their perception of the financial burden placed on them by their child’s cancer using a standard visual analogue scale (0, no burden to 100, greatest possible burden) in Table III. The mean perceived financial burden for the entire cohort was 66 (sd 32, median 73, IQR 50,87). Multivariable models showed a significantly greater burden in rural respondents (adjusted mean difference 9.6, 95% CI 0.3 to 18.8, p=0.04) and in those traveling >2 hours to the treatment center (adjusted mean difference 8.9, 95% CI 0.7 to 17.1, p=0.03).
Impact on Patient Education
Of caregivers of school-aged patients, 152 (60%) reported that their child was taken out of school during treatment. Of those continuing school, rural patients missed a more school days over the first six months from diagnosis than did urban patients (unadjusted means 90 vs. 58 days, respectively; adjusted mean difference 32.7, 95% CI 2.0 to 63.4, p=0.04). While there was no significant difference in the average number of school days missed between those traveling >1 hour or <1 hour to the treatment center (unadjusted means 78 vs. 57 days, respectively; adjusted mean difference 19.4, 95% CI −11.1 to 48.0, p=0.22), the difference was significant for those traveling >2 hours versus <2 hours (unadjusted means 87 vs. 57 days, respectively, p=0.05; adjusted mean difference 32.8, 95% CI 3.6 to 62, p=0.03). (Table III)
Even for patients staying in school, 28% of caregivers reported that their child was “not able to keep up with peers” and 10% reported repeating a grade, with rural patients being at an increased risk of repeating a grade (unadjusted percents: 16% of rural patients vs. 8% of urban patients; aRR 3.0, 95% CI 1.1 to 8.3, p=0.04). (Table II)
Impact of Diagnosis
The impact of patient diagnosis was also analyzed using the same outcomes of interest. Families of ALL patients were less likely to move because of cancer when compared to other diagnoses. Despite a small sample size, AML patients’ families showed a trend toward being more likely to move (RR 2.0, p=0.08) and were significantly more likely to move due to the cancer (RR 5.6, p=0.01). In addition, caregivers of AML patients were more likely to quit work when compared to those with other diagnoses (RR 2.6, p< 0.01). There were no other significant differences seen between diagnosis groups. (Supplementary Tables I and II)
Discussion
Our findings demonstrate a significant social and financial burden associated with caring for a child with cancer that, in many cases, is compounded by living remotely. While our institution serves a larger and less densely populated area than many pediatric oncology centers, most centers likely encounter similar challenges in delivering care to geographically underserved populations. Of note is that a threshold of a travel time of over two hours was consistently associated with a greater impact on patients and families.
For urgent care, rural residence was not as sensitive a predictor of emergency ground transport utilization as was the travel time from the treatment center. However, the proportion of patients requiring air-transport was significantly increased in both rural and remote patients, with nearly one-quarter of rural patients requiring this costly intervention (estimated $7,800–$25,000 per flight)19,20.
We found a surprisingly high percentage of caregivers who had moved residences because of their child’s cancer. Given the considerable effort and financial costs associated with moving, the high proportion of families undertaking such a major life change highlights the considerable impact a cancer diagnosis has on the lives of patients’ families. The trend towards having the move be more often “due to cancer” for those living further away suggests that the challenges encountered may be more pronounced for remote patients.
Our results indicate that having a child with cancer impacts caregivers’ employment. Rural and remote caregivers clearly missed more workdays for each clinic visit than did proximal patients. Interestingly, while remote caregivers missed more work than local caregivers during the first month after diagnosis, there was no significant difference in the amount of missed work between these groups when extended to include the first six months following the diagnosis. This finding suggests that many caregivers were potentially able to adjust to a new routine that accommodated their child’s treatment schedule. However, sacrifices or compromises made to accomplish this were not captured by this survey.
Perhaps the most striking impact on employment we observed was that over one-third of the cohort quit or changed jobs as a direct result of their child having cancer, with no differences based on location of residence. As we do not have data on the caregivers’ occupation, it is unclear if the amount of missed work or the risk of losing or changing jobs could be impacted by the nature or flexibility of a particular job type, which may differ between urban and rural populations. Also, it is likely that households with two working parents, a distinction not captured in this survey, would be impacted differently than those with a single wage earner.
We also found that the out-of-pocket expenses were substantially increased for remote families. As many patients require dozens of clinic visits over the course of therapy, a mean expense of $138 per visit for rural patients can quickly add up to a significant financial burden. Accordingly, the perception of financial burden imposed by the child’s cancer was significantly higher in rural families and those traveling greater than two hours. It is notable that the increased perception of financial burden in the remote patients exists despite an increased utilization of outside care in this population.
While patient diagnosis had no impact on most outcomes of interest, we found that AML had a greater impact on relocation and employment. This is likely due to the more intensive regimens and prolonged inpatient stays associated with AML therapy.
Our results are consistent with prior studies that have shown a significant financial burden for adult cancer patients and for caregivers of pediatric cancer patients15,21–23. However, to our knowledge, this study is among the first to document the impact of patients’ residence on the financial burden of childhood cancer in the United States. As most pediatric cancer centers serve a larger geographic area than do adult centers, a significant proportion of patients travel great distances for cancer care. Some of the burden on rural families might be moderated by reducing the number of visits to the tertiary care center. In light of our findings, further research on the outcomes, costs, and benefits of strategies such as tele-health, home health services, involving local primary care physicians, and reducing frequency of follow-up visits after completion of therapy is warranted.
Our study’s results also suggest a dramatic impact on schooling for the patients themselves. Over half elected to stop attending regular school while undergoing therapy. Of those continuing, rural patients missed more school days. While a travel time of >1 hour did not significantly impact the missed number of school days, traveling >2 hours did. The fact that 28% of patients had difficulty keeping up with their peers and 10% repeated a grade suggests that the demands of therapy significantly impact school performance, even for those continuing school. Our results also suggest the impact on schooling may be greater for rural patients as they more often repeated a grade. How childhood cancer’s impact on patient education translates to more long-term financial implications remains unstudied.
As is the case with employment, school performance and attendance is multifactorial, and our analysis does not account for potential differences in academic rigor, supportive resources, or student populations among rural and urban schools. Much of the existing data regarding the impact of cancer therapy on education focuses on the effect of the treatment on learning ability, concentration, memory, and comprehension 24. This underscores the need for more research on the impact of missed school and the development of interventions that maximize learning for children receiving complex cancer care. In particular, newer technologies, such as online video conferencing and internet-based curriculums may be helpful to students, particularly those from remote populations.
Our study has several limitations. Sample size is a limiting factor, which precludes evaluating small subset of patients. A larger multi-institutional study would allow for a better understanding of the burden experienced by patients from extremely isolated communities. Related, this study only captures patients who were ultimately treated at a tertiary hospital, perhaps underestimating the challenges for patients without access to such an institution. Also, the study population is heterogeneous with respect to patient diagnoses, which could impact the results. However, our study aimed to explore outcomes that are common to all caregivers of pediatric oncology patients and all multivariable models included diagnosis to minimize the potential for confounding.
While our institution follows standard protocols common to most institutions, this sample is from a single institution and may not fully represent all patients. In addition, the survey was administered only to caregivers of patients who were alive at the time of measurement, which may not accurately portray the burden for those whose child succumbed to their disease.
Participants were not selected based on phase of treatment, which could influence responses, although many questions focused on the first six months from diagnosis to facilitate comparisons. While our regression models included time since diagnosis, future longitudinal work should assess these outcomes at specified time points to better understand how treatment phase may influence burden. While over 98% of surveys were completed by a parent, we did not assess whether it was the mother or father, which could have offered additional valuable information. Finally, with the exception of a significant proportion of Hispanic respondents, our survey did not capture the experiences of other minority groups.
In summary, our data show that having a child with cancer places a large burden on families, as is evidence by missed work, changes in employment, relocation of residences, travel expenses and missed school. These burdens were greater for remote patients. Our study highlights issues unique to this underserved population and suggests the need for interventions that minimize the burden in such patients.
Supplementary Material
Acknowledgments
We thank the following for their support: University of Utah’s Center for Clinical and Translational Science for providing support, with funding from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, grant 8UL1TR000105 and the Huntsman Cancer Institute’s Cancer Center Support Grant supported by the National Cancer Institute, P30CA042014. Most importantly, we thank the patients and families for their participation.
Footnotes
Conflict of Interest Statement: There are no conflicts of interest between the authors and the subject matter of this manuscript.
References
- 1.Commerce. USDo: US Census Bureau, 2010 Census. , 2010.
- 2.Onega T, Duell EJ, Shi X, Wang D, et al. Geographic access to cancer care in the U.S. Cancer. 2008;112:909–18. doi: 10.1002/cncr.23229. [DOI] [PubMed] [Google Scholar]
- 3.Heath JA, Lintuuran RM, Rigguto G, Tokatlian N, et al. Childhood cancer: its impact and financial costs for Australian families. Pediatr Hematol Oncol. 2006;23:439–48. doi: 10.1080/08880010600692526. [DOI] [PubMed] [Google Scholar]
- 4.Lansky SB, Cairns NU, Clark GM, Lowman J, et al. Childhood cancer: nonmedical costs of the illness. Cancer. 1979;43:403–8. doi: 10.1002/1097-0142(197901)43:1<403::aid-cncr2820430157>3.0.co;2-l. [DOI] [PubMed] [Google Scholar]
- 5.Cohn RJ, Goodenough B, Foreman T, Suneson J. Hidden financial costs in treatment for childhood cancer: an Australian study of lifestyle implications for families absorbing out-of-pocket expenses. J Pediatr Hematol Oncol. 2003;25:854–63. doi: 10.1097/00043426-200311000-00007. [DOI] [PubMed] [Google Scholar]
- 6.Bodkin CM, Pigott TJ, Mann JR. Financial burden of childhood cancer. Br Med J (Clin Res Ed) 1982;284:1542–4. doi: 10.1136/bmj.284.6328.1542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dockerty JD, Skegg DC, Williams SM. Economic effects of childhood cancer on families. J Paediatr Child Health. 2003;39:254–8. doi: 10.1046/j.1440-1754.2003.00138.x. [DOI] [PubMed] [Google Scholar]
- 8.Miedema B, Easley J, Fortin P, Hamilton R, et al. The economic impact on families when a child is diagnosed with cancer. Curr Oncol. 2008;15:173–8. doi: 10.3747/co.v15i4.260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fletcher PC. My child has cancer: the costs of mothers' experiences of having a child with pediatric cancer. Issues Compr Pediatr Nurs. 2010;33:164–84. doi: 10.3109/01460862.2010.498698. [DOI] [PubMed] [Google Scholar]
- 10.Daniel G, Wakefield CE, Ryan B, Fleming CA, et al. Accommodation in pediatric oncology: parental experiences, preferences and unmet needs. Rural Remote Health. 2013;13:2005. [PubMed] [Google Scholar]
- 11.Scott-Findlay S, Chalmers K. Rural families' perspectives on having a child with cancer. J Pediatr Oncol Nurs. 2001;18:205–216. doi: 10.1053/jpon.2001.26864. [DOI] [PubMed] [Google Scholar]
- 12.Howat A, Veitch C, Cairns W. A descriptive study comparing health attitudes of urban and rural oncology patients. Rural Remote Health. 2006;6:563. [PubMed] [Google Scholar]
- 13.Smith SM, Campbell NC. Provision of oncology services in remote rural areas: a Scottish perspective. Eur J Cancer Care (Engl) 2004;13:185–92. doi: 10.1111/j.1365-2354.2003.00472.x. [DOI] [PubMed] [Google Scholar]
- 14.Samardakiewicz M, Kowalczyk JR. Psychosocial functioning of childhood cancer survivors living in rural areas. Ann Agric Environ Med. 2007;14:335–9. [PubMed] [Google Scholar]
- 15.Tsimicalis A, Stevens B, Ungar WJ, McKeever P, et al. A mixed method approach to describe the out-of-pocket expenses incurred by families of children with cancer. Pediatric blood & cancer. 2013;60:438–45. doi: 10.1002/pbc.24324. [DOI] [PubMed] [Google Scholar]
- 16. [Accessed September 30, 2013];Rural-Urban Commuting Area Codes (RUCAs) http://depts.washington.edu/uwruca/
- 17.United States Department of Agriculture. [Accessed September 30, 2013];Rural-Urban Commuting Codes. http://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes.aspx-.UgUbbm0kAVC.
- 18.Woodward M. Epidemiology: Study Design and Data Analysis, Chapman and Hall Texts in Statistical Science. New York: Chapman and Hall; 1999. pp. 459–460. [Google Scholar]
- 19.Delgado MK, Staudenmayer KL, Wang NE, Spain DA, et al. Cost-effectiveness of helicopter versus ground emergency medical services for trauma scene transport in the United States. Annals of emergency medicine. 2013;62:351–364 e19. doi: 10.1016/j.annemergmed.2013.02.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Aleccia J. Air ambulance leave some sky-high bills: Costs range form $12,000 to $25,000 a flight -and insurance may not pay. 2012 NBCNews.com.
- 21.Zafar SY, Peppercorn JM, Schrag D, Taylor DH, et al. The financial toxicity of cancer treatment: a pilot study assessing out-of-pocket expenses and the insured cancer patient's experience. The oncologist. 2013;18:381–90. doi: 10.1634/theoncologist.2012-0279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bernard DS, Farr SL, Fang Z. National estimates of out-of-pocket health care expenditure burdens among nonelderly adults with cancer: 2001 to 2008. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29:2821–6. doi: 10.1200/JCO.2010.33.0522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Langa KM, Fendrick AM, Chernew ME, Kabeto MU, et al. Out-of-pocket health-care expenditures among older Americans with cancer. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research. 2004;7:186–94. doi: 10.1111/j.1524-4733.2004.72334.x. [DOI] [PubMed] [Google Scholar]
- 24.Peckham VC, Meadows AT, Bartel N, Marrero O. Educational late effects in long-term survivors of childhood acute lymphocytic leukemia. Pediatrics. 1988;81:127–33. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.