Quality of prostate cancer care among rural men in the Veterans Health Administration
This article is US. Government work and, as such, is in the public domain in the United States of America.
Patient travel distances, coupled with variation in facility-level resources, create barriers for prostate cancer care in the Veterans Health Administration integrated delivery system. For these reasons, the authors investigated the degree to which these barriers impact the quality of prostate cancer care.
The Veterans Affairs Central Cancer Registry was used to identify all men who were diagnosed with prostate cancer in 2008. Patient residence was characterized using Rural Urban Commuting Area codes. The authors then examined whether rural residence, compared with urban residence, was associated with less access to cancer-related resources and worse quality of care for 5 prostate cancer quality measures.
Approximately 25% of the 11,368 patients who were diagnosed with prostate cancer in 2008 lived in either a rural area or a large town. Rural patients tended to be white (62% urban vs 86% rural) and married (47% urban vs 63% rural), and they tended to have slightly higher incomes (all P < .01) but similar tumor grade (P = .23) and stage (P = .12) compared with urban patients. Rural patients were significantly less likely to be treated at facilities with comprehensive cancer resources, although they received a similar or better quality of care for 4 of the 5 prostate cancer quality measures. The time to prostate cancer treatment was similar (rural patients vs urban patients, 96.6 days vs 105.7 days).
Rural patients with prostate cancer had less access to comprehensive oncology resources, although they received a similar quality of care, compared with their urban counterparts in the Veterans Health Administration integrated delivery system. A better understanding of the degree to which facility factors contribute to the quality of cancer care may assist other organizations involved in rural health care delivery. Cancer 2013;119:3629–3635. © 2013 American Cancer Society.
Prostate cancer is the most common cancer diagnosis and cause of cancer death among male US veterans. Because of the proportions of older men and African Americans served, as well as the increased prevalence of high-risk disease, a disproportionate prostate cancer burden is placed on the Veterans Health Administration (VA). Consequently, access to prostate cancer diagnosis and treatment resources is critical for this integrated delivery system.
However, travel distances, particularly for rural patients, coupled with variation in prostate cancer resources across urban and rural facilities because of regionalization, may create entry and structural access barriers to quality care.[2-4] Many veterans face significant transportation challenges, especially those residing in rural areas, creating barriers to entry for cancer care.[5, 6] These challenges may manifest as decreased compliance with care and can be exacerbated when it comes to seeking specialty care available at a limited number of facilities.[4, 7-9] Although the VA system has made significant efforts to ensure that its patients receive quality care no matter where they reside (eg, transportation to and from major medical centers, travel reimbursement, overnight lodging), the degree to which these efforts overcome the barriers that face rural patients with prostate cancer is unknown. More broadly, the problems facing rural veterans with prostate cancer may mirror those faced by other Americans living in rural areas. Respective solutions are increasingly relevant as accountable care organizations understand how best to care for their rural populations.[12, 13]
For these reasons, we investigated the degree to which entry and structural access barriers impact the quality of prostate cancer care for rural patients in the VA integrated delivery system. We hypothesized that rural patients treated at facilities with fewer cancer-related resources would receive lower quality prostate cancer care compared with their urban counterparts.
MATERIALS AND METHODS
We identified all patients who were diagnosed with pathologically confirmed, incident prostate cancer during the calendar year 2008 through the VA Central Cancer Registry. Diagnoses were confirmed through a national review of patients' medical records. We excluded patients who had no confirmation of prostate cancer in the medical record, a recurrence of prior prostate cancer (ie, not an incident case of prostate cancer), a diagnosis at autopsy, death ≤30 days after cancer diagnosis, an enrollment in hospice <30 days after diagnosis, metastatic cancer other than prostate, and a life expectancy ≤6 months documented in the medical record. Disease risk was classified into low, intermediate, and high risk according to D'Amico criteria. For this study, we linked the cancer registry data to corresponding VA administrative files, including inpatient, outpatient, laboratory, radiology, pharmacy, and facility data for 11,368 patients.
We hypothesized that patient residence and travel distance to their VA prostate cancer facility would pose barriers to quality cancer care. For example, we expected worse compliance with recommended radiation therapy treatment doses (ie, >75 grays [Gy]) for patients residing in rural areas because of transportation issues and the 8-week duration of treatment. We used a 4-tiered consolidation of Rural Urban Commuting Area (RUCA) codes to characterize the residence of men with prostate cancer as urban, suburban, large town, or rural according to their zip code. We further stratified urban and rural status for each patient according to the travel distance to their diagnosing facility (<12 miles vs >12 miles for urban, <30 miles vs >30 miles for rural). We hypothesized that these travel distance subcategories might correspond to those distances that would be expected to limit health care access within a respective RUCA classification. We used ArcGIS software (ESRI, Redlands, Calif) to compute these linear distances from the center of each patient's zip code to the center of their corresponding VA facility's zip code. This linear distance, ie, as the crow flies, is correlated with travel times. Moreover, nearly 90% of patients in this study received prostate cancer care at their diagnosing facility.
Facility Resources for Cancer Care
On the basis of a 2005 VA facility survey of oncology resources (138 facilities), we characterized each facility according to the following characteristics that might influence selected measures of quality prostate cancer care: American College of Surgeons Commission on Cancer certification; onsite cancer registrar; prostate cancer cases discussed at a tumor board; urologist, medical, and radiation oncology staffing; onsite prostate pathology services, chemotherapy, radiation therapy, and palliative care services. In total, 9862 patients had both residence (ie, RUCA) and facility level data available for this study.
Quality of Prostate Cancer Care
We selected 5 quality measures for prostate cancer care ascertained at the patient level through chart review based on their feasibility, validity, and importance to improving care. In most cases, recommended measures reflected evidence-based clinical guideline recommendations and were adapted from previously developed quality measures.[18-20] These included: 1) at least 10 core samples taken at prostate needle biopsy, 2) no bone scan before primary therapy for prostate cancer at low risk of recurrence, 3) central axis external-beam radiation therapy doses of at least 75 Gy, 4) administration or documented discussion of docetaxel-based chemotherapy for castration-resistant prostate cancer, and 5) neoadjuvant and/or adjuvant hormone therapy for high-risk patients receiving external-beam radiation therapy.[25, 26] We defined men who had documented signs or symptoms of progressive disease (eg, rising prostate-specific antigen, imaging results) during androgen-deprivation therapy as having castrate-resistant prostate cancer. There were 120 patients eligible nationally for this quality measure. Brachytherapy and cryotherapy were received by only 8% and <1% of patients, respectively, and were not separately examined.
Data used to calculate these indicators were collected remotely from the electronic medical record by trained medical abstractors using a VA contractor (West Virginia Medical Institute, Charlestown, WVa). Any discrepancies or questions were directed to local providers for clarification. The overall pass rates were based on care at any VA facility or paid for by the VA.
To better understand whether clinical characteristics or prostate cancer treatment varied according to patient residence (ie, urban, suburban, large town, or rural), we used the Mantel-Haenszel chi-square test and analysis of variance for categorical and continuous data, respectively. Next, we used chi-square testing and descriptive statistics to examine the relation between patient residence and facility resources for prostate cancer care, expecting that rural patients would be less likely to receive care at facilities that were rich with oncology resources. We also examined whether patient residence appeared to be associated with differences in average performance on any of the 5 quality measures using the Student t test. Last, we performed a subset analysis of patients with high-risk prostate cancer to examine their quality of care across residency status.
All analyses were performed using computerized software (SAS Institute, Cary, NC), and all testing was 2-sided. The probability of a Type I error was set at 0.05. The study protocol was approved by the VA research and development and institutional review boards of the VA Greater Los Angeles health care system.
Approximately 25% of patients diagnosed with prostate cancer in 2008 lived in either a rural area or a large town, and the majority of newly diagnosed veterans lived in an urban area, as indicated in Table 1. Not surprisingly, distance to the facility where patients received prostate cancer care increased with increasingly rural residence (urban vs rural, 15 miles vs 82 miles; P < .001). Rural patients tended to be white (urban vs rural, 62% vs 86%; P < .001), to be married (urban vs rural, 47% vs 63%; P < .001), and to have slightly higher incomes (P = .002). However, there were no significant differences in tumor grade or stage, and there were minimal differences with respect to initial treatment and comorbidity among rural patients compared with urban patients.
Table 1. Patient and Disease Characteristics According to Veteran Residence
|No. of patients (%)||7199 (63.3)||1045 (9.2)||1317 (11.6)||1772 (15.6)|| |
|Median distance to treating facility, miles||15||40||74||82||< .001|
|Age, %|| || || || ||< .001|
|<55 y||7||6||6||4|| |
|55–64 y||45||44||47||44|| |
|65–75 y||33||37||33||37|| |
|>75 y||15||13||14||15|| |
|Race/ethnicity, %|| || || || ||< .001|
|Income in US dollars, %|| || || || ||.002|
|Married, %||47||60||58||63||< .001|
|Comorbidity|| || || || ||.028|
|Tumor grade, %|| || || || ||.232|
|Well differentiated: Gleason score 2–4||1||1||1||1|| |
|Moderately differentiated: Gleason score 5–6||45||42||43||43|| |
|Poorly differentiated: Gleason score 7–10||53||54||54||54|| |
|Tumor classification, %|| || || || ||.119|
|Initial treatment, %|| || || || ||< .001|
|Androgen deprivation||16||19||14||18|| |
Nearly 60% of veterans with newly diagnosed prostate cancer were treated at facilities certified by the American College of Surgeons Commission on Cancer (Table 2). Urban patients were more likely to be treated at these facilities (urban vs rural, 63% vs 46%; P < .001). On average, rural patients were less likely to have medical oncologist, radiation oncologist, and urologist staffing at their treating facility compared with urban patients. Although the availability of palliative care resources was similar among urban and rural patients (P = .192), rural patients were less likely to attend facilities with radiation therapy services (P < .001).
Table 2. Oncology-Specific Facility Resources for Veteran Patients With Prostate Cancer Diagnosed in 2008
|American College of Surgeons Commission on Cancer?||50||59||63||56||49||46||<.001|
|On-site cancer registrar?||18||87||90||89||82||80||<.001|
|Prostate cancer cases discussed at a tumor board?||50||41||37||47||46||44||<.001|
|Specialists available|| || || || || || || |
|Pathology services|| || || || || || || |
|Prostate cancer pathologist on-site||31||42||47||35||34||31||<.001|
|Radiation therapy services||28||41||45||36||27||25||<.001|
|Palliative care services||91||93||93||94||94||92||.192|
Despite the differential access to structural resources for prostate cancer care, rural patients received similar or better quality of care for 4 of the 5 measures (Table 3), ie, 10 cores at prostate biopsy (P = .31), no bone scan for low-risk disease (P < .01), neoadjuvant androgen therapy for high-risk patients who received radiation therapy (P = .36), and docetaxel-based chemotherapy for castrate-resistant metastatic prostate cancer (P = .89). For example, nearly 90% of rural patients with low-risk disease did not receive an unnecessary bone scan for prostate cancer staging compared with 81% of urban patients.
Table 3. Quality of Prostate Cancer Care According to Veteran Residence and Travel Distance to Treatment Facility
|At least 10 core-needle samples taken at time of prostate biopsy||90.1 (30)||88.7 (31.6)||.31||89.7 (30.5)||90.8 (28.9)||.09||85.4 (35.4)||90.1 (29.8)||.01|
|No bone scan or PET scan before primary therapy for low-risk prostate cancer||80.9 (39.3)||88.9 (31.6)||<.01||84.6 (36.1)||80.2 (39.9)||.05||93.8 (24.4)||87.5 (33.1)||.23|
|Central axis doses of at least 75 Gy for radiotherapy||74.5 (43.6)||63.3 (48.5)||.01||74 (44)||70 (45.9)||.32||89 (31)||66 (47.9)||.01|
|Docetaxel-based chemotherapy for castration-resistant, metastatic prostate cancer||53.1 (50.3)||41.2 (50.7)||.89||52.9 (50.4)||64.3 (48.5)||.28||50 (70.7)||40 (50.7)||.80|
|Neoadjuvant and/or adjuvant hormone therapy for high-risk patients receiving external-beam radiotherapy||85.3 (35.4)||90.3 (29.7)||.36||82.1 (38.4)||87.1 (33.6)||.11||81.3 (40.3)||92.2 (27)||.18|
However, rural patients were less likely to be treated with full doses of radiation therapy (P = .01). When we further stratified rural and urban patients according to those distances that would be expected to limit health care access within a respective RUCA classification, the most dramatic difference occurred in the >75 Gy radiation dose measure for rural patients. Rural patients residing <30 miles from their treatment facility had compliance of nearly 90% compared with those residing >30 miles from their treatment facility, who received appropriate dosing 66% of the time (89% vs 66%; P = .01). The time to prostate cancer treatment for all patients was similar (rural vs urban, 96.6 days vs 105.7 days). In our subset analysis of patients with high-risk prostate cancer, we observed that rural patients received treatment in a similar timeframe as urban patients (rural vs urban, 106.6 days vs 107.9 days) and received similar degrees of neoadjuvant and/or adjuvant hormone therapy when treated with external-beam radiation therapy.
One in 4 veterans diagnosed with prostate cancer in 2008 lived in either a rural area or a large town. Rural patients with prostate cancer in the VA traveled nearly 5-fold farther for care and were significantly less likely to be treated at facilities that had comprehensive cancer resources compared with their urban counterparts. In contradiction to our hypothesis, however, this did not translate into worse quality of care for these patients, with the exception of slightly lower rates of radiation therapy completion. Rural patients actually fared better than their urban counterparts on an overuse measure of unnecessary imaging to stage newly diagnosed, low-risk disease. These findings suggest that, for the most part, the VA has mitigated entry and structural access barriers facing its rural patients with prostate cancer.
Increasing distance to a treatment facility can be a significant barrier to timely receipt of specialty care for any patient. Not surprisingly, travel distance for veterans with prostate cancer increased with increasing rural status. However, the average time to treatment did not vary across patients in rural versus urban areas despite rural patients traveling more than 80 miles on average for their prostate cancer care. This is particularly remarkable given a White River Junction VA Medical Center survey, which indicated that nearly 25% of veterans used friends, family, or the Disabled American Veterans van service to get to their appointments. This again suggests that VA efforts to ensure that veterans without access to personal transportation and who live in remote areas still have access to health care have been quite successful.[10, 28]
It is also interesting to note that rural patients in this equal-access, integrated delivery system did not harbor more aggressive prostate cancer (perhaps because of delay in diagnosis) and received timely, high-quality care most of the time. When we examined patients with perhaps the most to gain or lose from quality care (ie, those with high-risk prostate cancer), we observed that the care for rural and urban high-risk patients was similar. These findings suggest that, for patients with high-risk prostate cancer, whether they resided in a rural area or an urban area had limited impact on their quality of care in this integrated delivery system. Moreover, the finding that rural patients who were receiving care at facilities that had fewer structural indicators of quality were able to pass the panel indicates that a better understanding is needed of how facility factors contribute to the quality of cancer care and outcomes for veterans.
These findings highlight the reality that less than half of the facilities providing prostate cancer care are staffed by radiation oncologists or have the ability to deliver radiation therapy in the VA. Rural patients who lived greater than 30 miles away from their treatment facility struggled to comply with the prolonged treatment regimen, consistent with other literature examining radiation therapy delivery to rural patients with cancer. This is mitigated at least in part by expensive, fee-basis outsourcing strategies, yet it may compromise the quality of care in at least 2 ways. First, fee-basis strategies promote fragmentation of care and could lead to inefficient and even duplicate services. Second, the VA is unable to control whether contracted radiation oncologists deliver appropriate radiation therapy (eg, dose), potentially compromising quality of care. Although patients still may achieve a clinical response with slightly lower radiation doses, we did not examine the degrees of under-treatment across residency status. Indeed, radiation oncologists and their services are central to quality cancer care, and patients might benefit from further streamlining these resources in the VA.
In its 2003 Directive 2003-034, VA National Cancer Strategy, the VA stressed the importance of ensuring quality cancer care, facilitating continuing care, and enhancing access to care for all veterans with cancer. This directive was especially relevant to prostate cancer, because it identified the condition as service-connected for those exposed to the herbicide Agent Orange. The quality of prostate cancer care for rural and urban veterans with prostate cancer in this study was high, similar to reports from other studies that examined the quality of cancer care within and outside of the VA.[30, 33] However, further disentangling how and which patient and facility factors contribute to quality outcomes for veterans with cancer is warranted. Qualitative methods across a range of facilities with respect to cancer resources and care coordination are likely the next step toward informing scalable interventions to help all veterans get optimal prostate cancer care.
Prostate cancer quality-of-care measures may be perceived as relatively straightforward, perhaps limiting our ability to discern more granular, but significant variation in quality. However, several points should be noted. First, the measures in this study span the cancer care continuum from diagnosis and treatment to follow-up care and consistently demonstrate quality care for rural patients. Second, the VA health system is comprised of >90% male patients, making performance on these measures particularly relevant. If there were suggestions of poor-quality care, particularly for rural patients, then further efforts would need to be directed toward access (eg, the VA Office of Rural Health). The VA investment in transportation and, currently, in telemedicine may help guide other organizations to care for their rural populations. Last, although the patients in this study were treated at facilities that had fewer oncology resources, they received a similar quality of care compared with their urban counterparts. This finding suggests that research is warranted to gain a better understanding of the degree to which oncology-specific structural factors contribute to the quality of cancer care.
Limitations to our study warrant discussion. Other barriers to quality prostate cancer care that were not considered in this study included mental and physical challenges, which may compromise access and quality of care for veterans with prostate cancer. Nevertheless, most patients in this study met each of the quality indicators, demonstrating that these issues do not necessarily impede quality care. Furthermore, understanding how a facility's academic affiliation impacts quality of care as the largest provider of health care training in the United States is also an area of future research. Given its preference-sensitive nature and the limited number of eligible patients (n = 120), there was marginal compliance with the docetaxel-based measure that may warrant further study.
We did not have detailed transportation information (eg, reimbursement) for these patients., Many veterans use VA transportation resources to overcome travel barriers, as discussed previously. Because this system investigates innovative telemedicine approaches to improve access to specialty care, rural patients with prostate cancer may stand to benefit even further by having to travel less without compromising care.[34-36] Last, our quality measure panel consisted mainly of process-of-care measures. Although several measures had direct links to patient-centered outcomes (eg, improved survival), portfolios of quality measures for prostate cancer care are limited, and further work is needed to develop a wider range of prostate cancer quality-of-care metrics. At the time of this study, major quality organizations had even fewer measures than were available in the VA. Finally, establishing the value of these selected measures in relation to outcomes remains the work of future studies.
Most veterans receive high-quality prostate cancer care according to the 5 quality measures examined in the current study. Rural patients with prostate cancer have less access to comprehensive oncology resources; however, they receive a similar quality of care compared with their urban counterparts in the VA integrated delivery system. Better understanding the degree to which facility factors contribute to the quality of cancer care may assist other organizations involved in rural health care delivery.
This work was supported by the Veterans Health Administration Office of Analytics and Business Intelligence “Supportive Care, End of Life Care, and Crosscutting Quality Indicators for Patients with Cancer Within VA” Project. Dr. Skolarus is supported by a VA HSR&D Career Development Award - 2 (CDA 12-171).
CONFLICT OF INTEREST DISCLOSURES
Dr. Shelton is employed by the Veterans Administration (VA) and has received a grant from the University of California at Los Angeles for bladder cancer. Dr. Malin has received a grant from the Veterans Health Administration Office of Quality and Performance. At the time of the study, she was a full-time employee of the VA and is currently a staff member without compensation of the VA West Los Angeles Healthcare Center. She currently employed by WellPoint Inc. and owns stock in the company. She is also a member of the American Society of Clinical Oncology Quality of Care Committee. Dr. Saigal has received a grant from the RAND Corporation. Dr. Skolarus is a consultant for ArborMetrix, Inc., a healthcare analytics and software firm specializing in hospital- and specialty-based care, to support patient-centered prostate cancer projects.