Abstract
Purpose
To investigate provider specialty, care coordination, and cancer survivors’ comorbid condition care.
Methods
This retrospective cross-sectional Surveillance, Epidemiology and End Results (SEER)-Medicare study included cancer survivors diagnosed in 2004, 2–3 years post-cancer-diagnosis, in fee-for-service Medicare. We examined (1) provider specialties (primary care [PCPs], oncology specialists, other specialists) visited post-hospitalization; (2) role of provider specialties in chronic and acute condition management; and (3) an ambulatory care coordination measure. Outcome measures covered (1) visits post-hospitalization for nine conditions; (2) chronic disease management (lipid profile, diabetic eye exam, diabetic monitoring); (3) acute condition management (electrocardiogram [EKG] for congestive heart failure [CHF], imaging for CHF, EKG for transient ischemic attack, cholecystectomy, hip fracture repair).
Results
Among 8661 cancer survivors, patients were more likely to visit PCPs than oncologists or other specialists following hospitalizations for 8/9 conditions. Patients visiting a PCP (vs. not) were more likely to receive recommended care for 3/3 chronic and 1/5 acute condition indicators. Patients visiting an other specialist (vs. not) were more likely to receive recommended care for 3/3 chronic and 2/5 acute condition indicators. Patients visiting an oncology specialist (vs. not) were more likely to receive recommended care on 2/3 chronic indicators and less likely to receive recommended care on 1/5 acute indicators. Patients at greatest risk for poor coordination were more likely to receive appropriate care on 4/6 indicators.
Conclusions
PCPs are central to cancer survivors’ non-cancer comorbid condition care quality.
Implications for Cancer Survivors
PCP involvement in cancer survivors’ care should be promoted.
Keywords: cancer survivorship, comorbidity, quality of care
INTRODUCTION
The estimated number of cancer survivors in the US is growing rapidly, from 10 million in 2001 to 12 million in 2007 and almost 14 million by 2012 [1–2]. These cancer survivors require surveillance for recurrence, general primary and preventive care, and often, care for comorbid conditions [3], as well as having psychosocial issues addressed [3–5]. Previous research investigating care in cancer survivors has consistently shown that patients visiting a primary care provider (PCP) are most likely to receive recommended preventive care [6–12]. A study in the UK suggests an important role for the “robust primary care system” in providing quality care in long-term cancer survivors [13]. Research has also shown that PCP visits are associated with recommended cancer surveillance [14]. However, approximately 20% of breast and colorectal cancer survivors did not see a PCP in the second year after their cancer diagnosis [8, 11]. Further, some cancer survivors do not recognize the value provided by a PCP’s participation in their care [15–17].
The role of different provider specialties in comorbid condition care for cancer survivors is even more complex and has received less attention. Because cancer is particularly common in older individuals, with 60% of cancer survivors over age 65 [2], many cancer survivors also have comorbid conditions. In a population-based sample, 58% of cancer survivors had at least one comorbidity [18]. These comorbid conditions may have important implications for patient outcomes. Patients with certain early stage cancers (e.g., breast cancer) are more likely to die of causes other than cancer [19]. Further, the survival of breast cancer patients with comorbidity is equivalent to breast cancer patients with one stage worse disease without comorbidity (i.e., patients with stage 1 cancer with comorbidity survive at rates similar to patients with stage 2 cancer without comorbidity) [20].
The care for cancer survivors with multiple comorbid conditions is further complicated by the design of the US health care delivery system. In the US, the Medicare program is the insurer for individuals age 65 and over [21]. Traditionally, Medicare was a public insurer, which paid providers on a fee-for-service basis, covering basic health services, including inpatient care (Part A) and outpatient care (Part B, with an additional premium), subject to deductibles and co-insurance. Starting in 2006, outpatient prescription drug coverage was available (Part D), also with an additional premium through the traditional program. Because traditional Medicare pays providers on a fee-for-service basis, there is no requirement for a primary care provider ‘gatekeeper’ or, until recently, any payment for care coordination. Medicare Advantage (Part C) is an alternative to the traditional program and provides coverage through private health maintenance organizations, preferred provider organizations, or fee-for-service insurance, and may include additional benefits. In 2004, 87% of Medicare beneficiaries were covered by traditional Medicare, but by 2010, this number had decreased to 76% [22].
Given the prevalence and importance of comorbid conditions for cancer survivors and the complexities of the US health care delivery system, research investigating the quality of care cancer survivors receive for their non-cancer conditions is critical. While previous investigations have demonstrated the PCP’s central role in preventive care for older cancer survivors in traditional Medicare fee-for-service, the role of the PCP in providing comorbid condition care for cancer survivors is less clear. Because survivors with comorbid conditions may be seeing both cancer specialists and other specialists, it is unknown whether PCPs’ involvement promotes quality comorbid condition care or simply adds another provider to the mix. Thus, we investigated the association between comorbid condition care quality indicators, involvement by different provider specialties, and risk of poor care coordination in cancer survivors.
METHODS
Study Design and Research Questions
Using published indicators [23] applied in previous studies [7], this retrospective cross-sectional study investigated the quality of acute and chronic comorbid condition care for survivors of breast, prostate, or colorectal cancer during Years 2–3 post-cancer diagnosis (assuming that cancer treatment would occur during Year 1). We investigated (1) which provider specialties cancer survivors visited after hospitalizations for nine different conditions: depression, diabetes, hypertension, gastrointestinal bleeding, transient ischemic attack (TIA), cerebrovascular accident (CVA), angina, myocardial infarction, and congestive heart failure (CHF); (2) the association between provider specialties visited and 3 chronic (lipid profile, diabetic eye exam, diabetic monitoring) and 5 acute (electrocardiogram [EKG] for CHF, imaging for CHF, EKG for TIA, cholecystectomy, hip fracture repair) comorbid condition quality indicators; and (3) the association between an ambulatory care coordination risk measure and the 6 ambulatory comorbid condition quality indicators. The first two sets of analyses were descriptive. For the third, we hypothesized that patients at greatest risk for poor coordination would be less likely to receive appropriate comorbid condition care. The Johns Hopkins School of Medicine Institutional Review Board deemed this analysis exempt.
Data Sources
We used the Surveillance, Epidemiology and End Results (SEER)-Medicare linked database [24], which links the SEER population-based cancer registries with Medicare claims. At the time of this study, SEER registries covered approximately 25% of the US population. For provider specialty, the unique provider identification numbers from the SEER-Medicare dataset were linked with the American Medical Association (AMA) provider database by a third-party. For the care coordination risk measure, which uses Medicare specialty codes, we mapped the AMA physician specialty codes to the Medicare specialty codes.
Study Subjects
This analysis included cancer survivors diagnosed with loco-regional breast, prostate, or colorectal cancer in 2004 and followed through 2007. Survivors were ≥66 years old, survived ≥3 years, and continuously enrolled in fee-for-service Medicare (Parts A and B) from one year prior through three years following their cancer diagnosis. As noted above, in 2004, 87% of Medicare beneficiaries were covered by traditional fee-for-service Medicare [22]. Because we were interested in care for cancer survivors who had completed active treatment and had no evidence of disease, we excluded subjects who received chemotherapy, radiation, or hospice care during the survivorship period (i.e., Years 2–3 post-diagnosis).
Variables
Outcome Variables
The outcome variables were previously applied [7] indicators of quality comorbid condition care developed by RAND [23] (Box 1). We used two separate groups of indicators (see Study Design and Box 1). The first group of quality indicators addressed visits within specified time periods following discharge for patients who had been hospitalized for one of nine conditions. We examined incident hospitalizations during the survivorship period (Years 2–3). The second group of quality indicators addressed 3 care processes for chronic conditions and 5 care processes for acute conditions. For the three chronic condition indicators, we used the Medicare claims from Year 1 post-cancer-diagnosis to identify subjects with the relevant conditions who were therefore eligible for the process measures. For the acute care indicators, we examined incident events during Years 2–3 post-cancer-diagnosis.
Box 1. Quality Indicators [23].
Visits Following Hospitalizations
|
Chronic Care Indicators
|
Acute Care Indicators
|
Excluded from the coordination risk measure analysis because the coordination risk measure is limited to ambulatory conditions.
Independent Variables
Physician specialty was categorized as PCP, oncology specialist, and other specialist, using the same categories applied in previous studies [6–11]. PCPs included specialties such as family medicine, internal medicine, and obstetrics/gynecology. Oncology specialists included specialties such as oncologists, hematologists, cancer surgeons, and general surgeons. All other specialty types were categorized as other specialists. We used three dichotomous variables to indicate visits to each provider category (yes/no). This categorization enabled us to compare, for example, care received by patients with PCP visits to care received by patients without PCP visits.
We also examined the association between chronic and acute care receipt and the Adjusted Clinical Groups® (ACG) Coordination Markers™, which categorize the likelihood of ambulatory care coordination issues as Likely, Possible, or Unlikely [25]. The coordination risk measure factors in the number of unique providers, the number of specialties involved, the majority source of care visit percentage, and whether a generalist (PCP, physician assistant, or nurse practitioner) was seen. Patients with more unique providers, more specialties involved, lower percentage of care from the majority source, and no generalist visit are considered to be at greater risk for coordination issues. Only evaluation and management visits with providers who could be involved in the overall management of the patient’s care are counted. The coordination risk measure has demonstrated validity, with greater coordination risk associated with higher costs and higher rates of asthma emergency department visits [25].
Covariates
Multivariable models included age 75+ (vs. <75), race (non-white vs. white), cancer type (breast or prostate vs. colorectal), urban residence (vs. rural), state buy-in participation (vs. not), and pre-cancer-diagnosis comorbidity score 1+ (vs. 0) [26–28]. State buy-in reflects whether a Medicare beneficiary’s state was responsible for paying the Part B premium, thus providing an indication of socio-economic status. The categorical pre-cancer-diagnosis comorbidity score grouped patients into those with a comorbidity score of 0 versus 1 or greater. Eligibility for given quality indicators was determined using the specific criteria from the published quality indicators [23].
Analyses
After describing the sample characteristics, we performed three sets of analyses. First, we examined which provider specialties (PCP, oncology specialist, other specialist) patients visited following hospitalization for nine conditions. We included all physician specialties visited during the post-discharge period specified by the indicator (4 weeks for all, except 2 weeks for depression), so patients may be counted as having visits in multiple provider categories.
Second, we used logistic regression, adjusting for the covariates, to examine whether visits to the three provider specialties were associated with receipt of quality comorbid condition care. We developed separate models for each of the 8 quality indicators (3 chronic and 5 acute).
Third, we used logistic regression, adjusting for the covariates, to examine the association between the 3 chronic and 3 of the 5 acute care indicators and the care coordination risk measure. Two acute care indicators related to inpatient care (cholecystectomy and hip fracture repair) were not included in these analyses as the coordination risk measure addresses ambulatory care. Analyses used SAS 9.1.3.
RESULTS
The sample included 8661 cancer survivors, mean age 75 years, 65% male, and 85% White (Table 1). The majority of the sample had survived prostate cancer (53%), with 22% surviving breast cancer, and 26% colorectal cancer. The vast majority (90%) resided in an urban setting, and 10% ever participated in state buy-in (indicating lower socioeconomic status).
Table 1.
Sample Characteristics
| Characteristic | Cases Overall (n=8661) | Colorectal Cases (n=2231) | Breast Cases (n=1871) | Prostate Cases (n=4559) |
|---|---|---|---|---|
| Age | ||||
| Mean (SD) | 74.8 (6.55) | 77.2 (7.02) | 76.9 (7.16) | 72.8 (5.25) |
| Sex n(%) | ||||
| Male | 5614 (64.8) | 1055 (47.3) | 0 (0.0) | 4559 (100.0) |
| Race n(%) | ||||
| White | 7330 (84.6) | 1926 (86.3) | 1656 (88.5) | 3748 (82.2) |
| Black | 721 (8.3) | 149 (6.7) | 113 (6.0) | 459 (10.1) |
| Other | 610 (7.0) | 156 (7.0) | 102 (5.5) | 352 (7.7) |
| Comorbidity Score n(%)a | ||||
| 0 | 6114 (70.6) | 1376 (61.7) | 1306 (69.8) | 3432(75.3) |
| 1+ | 2547 (29.4) | 855 (38.3) | 565 (30.2) | 1127(24.7) |
| Ever in State Buy-In n(%)b | ||||
| Yes | 883 (10.2) | 280 (12.6) | 245 (13.1) | 358 (7.9) |
| Urban/Rural Residence n(%) | ||||
| Urban | 7824 (90.3) | 1963 (88.0) | 1679 (89.7) | 4182 (91.7) |
| SEER Region n(%) | ||||
| Connecticut | 621 (7.2) | 179 (8.0) | 171 (9.1) | 271 (5.9) |
| Detroit | 586 (6.8) | 156 (7.0) | 112 (6.0) | 318 (7.0) |
| Hawaii | 131 (1.5) | 37 (1.7) | 19 (1.0) | 75 (1.7) |
| Iowa | 620 (7.2) | 229 (10.3) | 163 (8.7) | 228 (5.0) |
| New Mexico | 215 (2.5) | 54 (2.4) | 41(2.2) | 120 (2.6) |
| Seattle | 633 (7.3) | 119 (5.3) | 130 (7.0) | 384 (8.4) |
| Utah | 356 (4.1) | 62 (2.8) | 56 (3.0) | 238 (5.2) |
| Atlanta & Rural Georgia | 233 (2.7) | 43 (1.9) | 50 (2.7) | 140 (3.1) |
| Californiac | 2905 (33.5) | 662 (29.7) | 581 (31.1) | 1662 (36.5) |
| Kentucky | 579 (6.7) | 177 (7.9) | 138 (7.4) | 264 (5.8) |
| Louisiana | 572 (6.6) | 118 (5.3) | 137 (7.3) | 317 (7.0) |
| New Jersey | 1210 (14.0) | 395 (17.7) | 273 (14.6) | 542 (11.9) |
The pre-cancer-diagnosis categorical comorbidity score grouped patients into those with a comorbidity score of 0 versus 1 or greater.
State buy-in reflects whether a Medicare beneficiary’s state was responsible for paying the beneficiary’s Part B premium, and thus provides an indication of socio-economic status.
Including Greater California, San Francisco, San Jose, and Los Angeles registries.
The analysis of provider specialties visited after hospitalizations found substantial PCP involvement (Figure 1). For 8 of 9 conditions, patients were most likely to visit a PCP, ranging from 67% for acute myocardial infarction (tying other specialists) to 85% for malignant/severe hypertension. The only condition for which patients were not most likely to see a PCP was unstable angina (39% visiting PCP vs. 85% visiting other specialists). Oncology specialists were infrequently visited following hospitalizations for these non-cancer conditions (<22% of patients).
Figure 1.
Percentage of patients visiting each of the three provider specialty types (primary care providers [PCPs], other specialists, and oncology specialists) following hospitalization for 9 different conditions. Time period for the visits is 4 weeks from discharge for all but the depression indicator, which is 2 weeks from discharge. (HTN=hypertension, GI=gastrointestinal, TIA=transient ischemic attack, CVA=cerebrovascular accident, AMI=acute myocardial infarction, CHF=congestive heart failure)
The second analysis examined the association between provider specialties visited and 3 chronic and 5 acute quality indicators (Figure 2a). Compared to patients who did not visit other specialists, patients with other specialist visits were more likely to receive recommended care on all three chronic and two acute (EKG for CHF, chest imaging for CHF) indicators. Compared to patients who did not visit PCPs, patients with PCP visits were more likely to receive recommended care on all chronic and one acute (chest imaging for CHF) indicators. Compared to patients who did not visit oncology specialists, patients with oncology specialist visits were more likely to receive recommended care on two chronic (diabetic eye exam, diabetes monitoring) but less likely to receive recommended care on one acute (EKG for TIA) indicators. No other associations between the quality indicators and provider specialty visited were statistically significant.
Figure 2.
(a) Odds ratios and 95% confidence intervals for the association between physician specialties visited and comorbid condition quality indicators. Odds ratios greater than 1.0 indicate greater odds of recommended care receipt compared to no visit to the provider type. (b) Odds ratios and 95% confidence intervals for the association between the coordination risk measure and comorbid condition quality indicators. Odds ratios greater than 1.0 indicate greater odds of recommended care receipt compared to the reference group (unlikely coordination risk). Chronic condition indicators are denoted with diamonds; acute condition indicators are denoted with squares. Models are adjusted for age (75+ vs. <75), race (Non-white vs. White), cancer type, comorbidity score (1+ vs 0), state buy-in, and urban residence (vs. rural). (EKG=electrocardiogram, CHF=congestive heart failure, TIA=transient ischemic attack)
The third analysis examined the association between the care coordination risk measure and the three chronic condition indicators and three acute condition indicators focused on ambulatory care (Figure 2b). Compared to patients with an unlikely coordination risk, patients with a likely coordination risk were more likely to receive recommended care on four indicators (lipid profile, diabetic eye exam, diabetic monitoring, EKG for CHF) and patients with a possible coordination risk were more likely to receive recommended care on diabetic eye exam. The findings did not support our hypothesis that patients with a greater coordination risk would be less likely to receive recommended care.
To explore this unexpected finding, we repeated the models using the individual components of the coordination risk measure. Specifically, we used logistic regression models with performance on each indicator (yes/no) as a function of unique provider count (1, 2–6, 7+, other [meaning no visits to an eligible provider]), specialty count (<5 or ≥5), majority source of care (≤28% or >28%), and generalist seen (yes/no) while adjusting for the same covariates. These post-hoc analyses demonstrated that having more unique providers was consistently associated with more appropriate care based on the indicators (Table 2). For example, compared to patients with 1 unique provider, patients with 7 or more unique providers were more likely to receive recommended care on four indicators (diabetic eye exam, diabetic monitoring, EKG for CHF, chest imaging for CHF). Because the coordination risk measure considers more unique providers to be a higher coordination risk and because the post-hoc analysis demonstrated that more unique providers was associated with more appropriate comorbid condition care, the findings contradicted our initial hypothesis that patients at greater coordination risk would be less likely to receive recommended care.
Table 2.
Post-Hoc Analyses of the Association between the Care Coordination Risk Measure Individual Components and Comorbid Condition Quality Indicators
| Quality Indicator | Odds Ratios (95% Confidence Intervals)a | |||||
|---|---|---|---|---|---|---|
| Unique Provider Count 2–6 (vs. ≥7) | Unique Provider Count 1 (vs. ≥7) | Unique Provider Count Other (vs. ≥7) | Specialty Count ≥5 (vs. <5) | Majority Source of Care>28% (vs. ≤28%) | Generalist Seen (Y vs. N) | |
| Lipid Profile | 0.77 (0.37, 1.63) | 0.46 (0.16, 1.30) | 0.30 (0.05, 1.75) | 1.38 (0.70, 2.72) | 0.67 (0.25, 1.76) | 2.91 (1.62, 5.21) |
| Eye Exam for Diabetics | 0.86 (0.63, 1.18) | 0.31 (0.19, 0.52) | 0.21 (0.10, 0.46) | 1.57 (1.20, 2.05) | 0.96 (0.65, 1.43) | 0.94 (0.70, 1.27) |
| Diabetic Monitoring | 0.82 (0.58, 1.16) | 0.42 (0.23, 0.77) | 0.42 (0.16, 1.13) | 1.27 (0.94, 1.72) | 1.48 (0.95, 2.31) | 1.18 (0.83, 1.69) |
| EKG for CHF | 0.90 (0.60, 1.35) | 0.48 (0.25, 0.92) | 0.48 (0.20, 1.15) | 1.09 (0.74, 1.59) | 0.70 (0.43, 1.15) | 1.17 (0.79, 1.73) |
| Chest Imaging for CHF | 0.63 (0.42, 0.96) | 0.42 (0.22, 0.82) | 0.55 (0.23, 1.31) | 0.65 (0.44, 0.96) | 0.74 (0.46, 1.20) | 1.24 (0.83, 1.84) |
| EKG for TIA | 0.89 (0.44, 1.78) | 1.40 (0.37, 5.29) | 1.79 (0.28, 11.55) | 0.79 (0.42, 1.50) | 1.16 (0.52, 2.59) | 2.17 (0.92, 5.13) |
Logistic regression models with performance on each indicator (yes/no) as a function of unique provider count (1, 2–6, 7+, other [meaning no visits to an eligible provider]), specialty count (<5 or ≥5), majority source of care (≤28% or >28%), and generalist seen (yes or no) while adjusting for age (75+ vs. <75), race (Non-white vs. White), cancer type, comorbidity score (1+ vs 0), state buy-in, and urban residence (vs. rural)
EKG=electrocardiogram, CHF=congestive heart failure, TIA=transient ischemic attack
BOLD=p<0.05
DISCUSSION
Many cancer survivors have comorbid conditions, but there has been little research investigating the quality of comorbid condition care they receive. Earle & Neville [7] examined comorbid condition care quality in five-year colorectal cancer survivors diagnosed in 1991–1992 and found that having both a PCP and oncology specialist involved in the survivor’s care was associated with higher quality. Our study builds on this previous research by investigating care in survivors of three different cancer types, using more recent data, and focusing on the critical period when cancer treatment is ending and patients are transitioning to survivorship. We not only examined which provider specialties are involved in care but also investigated a coordination risk measure.
There were two key findings from this research. First, in addition to the expected importance of other specialists’ involvement in cancer survivors’ comorbid condition care, PCPs also played a pivotal role. PCP and other specialist visits were associated with all three indicators of chronic comorbid condition care and one (chest imaging for CHF) and two (EKG for CHF, chest imaging for CHF) indicators of acute condition care, respectively. That the statistically significant findings were all in the same direction (and the non-significant estimates tended to follow this trend) suggests that these associations are not the result of random variation. Further, on almost all of the post-hospitalization visit indicators, patients were most likely to see a PCP. In contrast, the results regarding oncology specialists’ involvement were less consistent, suggesting no association with comorbid condition care quality. While it may be unsurprising that these non-cancer condition-specific indicators may not be addressed by oncologists, a previous study demonstrated that, for 21% of oncology patients, oncologists are providing more non-cancer-related care than cancer-related care [29]. Taken together, these findings suggest that certain specialty types may play important roles in specific areas but PCPs are critical across chronic and post-hospitalization comorbid condition care. Given that previous research has found differences in comorbid condition care quality by the type of cancer survived [30], interventions that promote PCP involvement in survivors’ follow-up care may be effective.
The second important finding from this research is the relationship between the care coordination risk measure and the quality indicators. While one might expect a higher risk of poor coordination to be associated with worse comorbid condition care, in this analysis, higher risk of poor coordination was associated with more appropriate care. This finding was explained by the role of the unique provider count. Patients with more unique providers and likely risk for coordination issues were more likely to receive recommended care. Upon closer examination, this finding is explained by the fact that several of the quality indicators used in this analysis require more unique providers to be involved (e.g., eye exam for diabetics). This would explain why patients assigned to the likely coordination risk group would be more, rather than less, likely to receive appropriate care. It is also possible that the severity of the comorbid condition could be confounding the association. Another consideration is that we used the same study period to assign patients to coordination risk category and to evaluate the quality indicators. A prospective analysis in which the coordination risk category is assigned based on claims prior to the study period may be more effective.
Notably, the measure used in our analysis to assess coordination of care is described as a metric of risk for poorly coordinated care. It was validated using cost and asthma emergency department visits [25]. Because cancer survivors with comorbid conditions may have a complex array of providers involved in their care, the motivation for this analysis was to explore whether coordination risk was associated with recommended care receipt. For the reasons described above, it did not perform effectively for this purpose. Alternative metrics that assess care coordination – rather than coordination risk – are needed to provide a more complete evaluation of care quality. As defined by McDonald et al., care coordination is “the deliberate organization of patient care activities between two or more participants (including the patient) involved in a patient’s care to facilitate the appropriate delivery of health care services” [31 p.41]. Because our analysis relied on claims data, we could only assess “the appropriate delivery of health care services” -- not the more complex aspects of coordination such as whether there was a “deliberate organization” of care. Supplementing these claims-based process measures by surveying patients and providers would provide a more complete picture of care coordination. Given the challenges of measuring care coordination, future research would benefit from developing improved metrics of care coordination and from using multiple metrics (e.g., patient and provider surveys and claims data) simultaneously.
The study’s limitations also inform interpretation of the findings. SEER-Medicare only includes the population over age 65 in fee-for-service Medicare. Although these quality indicators were developed specifically for use with claims data [23], claims analysis only captures whether covered services were billed to Medicare -- not why services were or were not provided. Further, these quality indicators are limited to selected process-of-care measures and do not address the important psychosocial elements of survivorship care. Notably, based on the categorical comorbidity score, 29% of our sample had at least one comorbidity. This claims-based estimate is not directly comparable to the 58% estimate based on self-report from the population-based sample of cancer survivors [18] because (1) it was assessed based on claims data prior to the cancer diagnosis to estimate pre-cancer comorbidity burden, rather than in cancer survivors and (2) used a different approach to measure comorbidity [32]. As noted in the Methods, eligibility for given comorbid condition quality indicators was determined using the specific criteria from the published quality indicators [23], not from the comorbidity score reported in Table 1. Further, our sample might represent a healthier population because it only included subjects who survived a full three years following their cancer diagnosis. Examination of care quality among survivors who died during follow-up would be informative but confounds the determination of eligibility for the quality indicators.
Despite these limitations, this analysis addresses the important issue of comorbid condition care in cancer survivors. SEER-Medicare enables analysis of a large population-based sample from geographically diverse regions of the US. That we used the AMA provider file to determine physician specialty is another strength, as the data in the AMA file contain more detail regarding provider specialty. These findings support the critical role of PCPs in providing comorbid condition care even for complex cancer survivors.
Acknowledgments
The project described was supported by Award Number R01CA149616 from the National Cancer Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health. The findings have been presented in part at the 2012 Annual Meeting of the Society for General Internal Medicine, the 2012 Annual Meeting of the American Society of Clinical Oncology, and the 2012 Biennial Cancer Survivorship Conference.
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
The collection of the California cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract N01-PC-35136 awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention’s National Program of Cancer Registries, under agreement #U55/CCR921930-02 awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California, Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors is not intended nor should be inferred.
Footnotes
Disclosures: Dr. Lemke is a member of a group of faculty and staff at The Johns Hopkins University who develop and maintain the Adjusted Clinical Group (ACG) method. The Johns Hopkins University holds the copyright to the ACG software. To help support research and development, The Johns Hopkins University receives royalties from health plans and other organizations that license the ACG software. Dr. Frick consulted for and sits on a medical advisory board for eviti, a company that provides a service to determine whether a recommended cancer treatment regiment is consistent with evidence as reported in the literature or other guidelines and is consistent with insurers’ payment plans. No other authors have relevant conflicts of interest to disclose.
References
- 1.American Cancer Society. Cancer Treatment & Survivorship Facts & Figures 2012–2013. Atlanta, GA: American Cancer Society; 2012. [Google Scholar]
- 2.Rowland JH, Mariotto A, Alfano CM, Pollack LA, Weir HK, White A. Cancer survivors – United States, 2007. MMWR. 2011;60:269–272. [PubMed] [Google Scholar]
- 3.Institute of Medicine. From Cancer Patient to Cancer Survivor: Lost in Transition. Washington, DC: National Academy Press; 2006. [Google Scholar]
- 4.Armes J, Crowe M, Colbourne L, et al. Patients’ supportive care needs beyond the end of cancer treatment: a prospective, longitudinal survey. J Clin Oncol. 2009;20;27:6172–6179. doi: 10.1200/JCO.2009.22.5151. [DOI] [PubMed] [Google Scholar]
- 5.Aaronson NK, Mattioli V, Minton O, et al. Beyond treatment – Psychosocial and behavioural issues in cancer survivorship research and practice. Eur J Cancer Suppl. 2014;12:54–64. doi: 10.1016/j.ejcsup.2014.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Earle CC, Burstein HJ, Winer EP, Weeks JC. Quality of non-breast cancer health maintenance among elderly breast cancer survivors. J Clin Oncol. 2003;21:1447–51. doi: 10.1200/JCO.2003.03.060. [DOI] [PubMed] [Google Scholar]
- 7.Earle CC, Neville BA. Under use of necessary care among cancer survivors. Cancer. 2004;101:1712–19. doi: 10.1002/cncr.20560. [DOI] [PubMed] [Google Scholar]
- 8.Snyder CF, Earle CC, Herbert RJ, Neville BA, Blackford AL, Frick KD. Trends in follow-up and preventive care for colorectal cancer survivors. J Gen Intern Med. 2008;23:254–9. doi: 10.1007/s11606-007-0497-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Snyder CF, Frick KD, Peairs KS, et al. Comparing care for breast cancer survivors to non-cancer controls: a five-year longitudinal study. J Gen Intern Med. 2009;24:469–474. doi: 10.1007/s11606-009-0903-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Snyder CF, Earle CC, Herbert RJ, Neville BA, Blackford AL, Frick KD. Preventive care for colorectal cancer survivors: a 5-year longitudinal study. J Clin Oncol. 2008;26:1073–9. doi: 10.1200/JCO.2007.11.9859. [DOI] [PubMed] [Google Scholar]
- 11.Snyder CF, Frick KD, Kantsiper ME, et al. Prevention, screening, and surveillance care for breast cancer survivors compared with controls: changes from 1998 to 2002. J Clin Oncol. 2009;27:1054–1061. doi: 10.1200/JCO.2008.18.0950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Snyder CF, Frick KD, Herbert RJ, et al. Preventive care in prostate cancer patients: following diagnosis and for five-year survivors. J Cancer Surviv. 2011;5:283–291. doi: 10.1007/s11764-011-0181-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Khan NF, Mant D, Rose PW. Quality of care for chronic diseases in a British cohort of long-term cancer survivors. Ann Fam Med. 2010;8:418–424. doi: 10.1370/afm.1162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Salz T, Weinberger M, Ayanian JZ, et al. Variation in use of surveillance colonoscopy among colorectal cancer survivors in the United States. BMC Health Services Research. 2010;10:256. doi: 10.1186/1472-6963-10-256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cheung WY, Neville BA, Cameron DB, Cook EF, Earle CC. Comparison of patient and physician expectations for cancer survivorship care. J Clin Oncol. 2009;27:2489–2495. doi: 10.1200/JCO.2008.20.3232. [DOI] [PubMed] [Google Scholar]
- 16.Hudson SV, Miller SM, Hemler J, et al. Adult cancer survivors discuss follow-up in primary care: ‘not what I want, but maybe what I need’. Ann Fam Med. 2012;10:418–427. doi: 10.1370/afm.1379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kantsiper M, McDonald EL, Geller G, Shockney L, Snyder C, Wolff AC. Transitioning to breast cancer survivorship: perspectives of patients, cancer specialists, and primary care providers. J Gen Intern Med. 2009;24:459–466. doi: 10.1007/s11606-009-1000-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yabroff KR, Lawrence WF, Clauser S, Davis WW, Brown ML. Burden of illness in cancer survivors: findings from a population-based sample. J Natl Cancer Inst. 2004;96:1322–1330. doi: 10.1093/jnci/djh255. [DOI] [PubMed] [Google Scholar]
- 19.Hanrahan EO, Gonzalez-Angulo AM, Giordano SH, et al. Overall survival and cause-specific mortality of patients with stage T1a,bN0M0 breast carcinoma. J Clin Oncol. 2007;25:4952–4960. doi: 10.1200/JCO.2006.08.0499. [DOI] [PubMed] [Google Scholar]
- 20.Patnaik JL, Byers T, DiGuiseppi C, Denberg TD, Dabelea D. The influence of comorbidities on overall survival among older women diagnosed with breast cancer. J Natl Cancer Inst. 2011;103:1101–1111. doi: 10.1093/jnci/djr188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kaiser Family Foundation. [Accessed February 4, 2015];Medicare: A Primer. 2007 https://kaiserfamilyfoundation.files.wordpress.com/2013/01/7615lp.pdf.
- 22.Kaiser Family Foundation. [Accessed February 4, 2015];Medicare: A Primer. 2010 https://kaiserfamilyfoundation.files.wordpress.com/2013/01/7615-03.pdf.
- 23.Asch SM, Sloss EM, Hogan C, Brook RH, Kravitz RL. Measuring underuse of necessary care among elderly Medicare beneficiaries using inpatient and outpatient claims. JAMA. 2000;284:2325–2333. doi: 10.1001/jama.284.18.2325. [DOI] [PubMed] [Google Scholar]
- 24.National Cancer Institute. [Accessed November 21, 2014];SEER-Medicare: Brief description of the SEER-Medicare database. http://healthservices.cancer.gov/seermedicare/overview/brief.html.
- 25.The Johns Hopkins ACG® System. [Accessed November 21, 2014];Technical Reference Guide. Version 10.0. http://acg.jhsph.org/public-docs/ACGv10.0TechRefGuide.pdf.
- 26.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40:373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 27.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 28.Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53:1258–1267. doi: 10.1016/s0895-4356(00)00256-0. [DOI] [PubMed] [Google Scholar]
- 29.Rosenblatt RA, Hart LG, Baldwin L-M, Chan L, Schneeweiss R. The generalist role of specialty physicians: Is there a hidden system of primary care? JAMA. 1998;279:1364–70. doi: 10.1001/jama.279.17.1364. [DOI] [PubMed] [Google Scholar]
- 30.Snyder CF, Frick KD, Herbert RJ, et al. Quality of care for comorbid conditions during the transition to survivorship: differences between cancer survivors and non-cancer controls. J Clin Oncol. 2013;31:1140–1148. doi: 10.1200/JCO.2012.43.0272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.McDonald KM, Sundaram V, Bravata DM, et al., editors. AHRQ Publication No. 04(07)-0051-7. Rockville, MD: Agency for Healthcare Research and Quality; Jun, 2007. Closing the Quality Gap: A Critical Analysis of Quality Improvement Strategies. Technical Review 9 (Prepared by the Stanford University-UCSF Evidence-based Practice Center under contract 290-02-0017) [Google Scholar]
- 32.Sarfati D. Review of methods used to measure comorbidity in cancer populations: no gold standard exists. J Clin Epidemiol. 2012;65:924–933. doi: 10.1016/j.jclinepi.2012.02.017. [DOI] [PubMed] [Google Scholar]