Pathways, Outcomes, and Costs in Colon Cancer: Retrospective Evaluations in 2 Distinct Databases

, , , , , , ,
Supplements and Featured Publications, Special Issue: Payer/Provider Relationships in Oncology , Volume 17, Issue 5 Suppl

Retrospective evaluations of electronic health records and claims databases to assess clinical outcomes and costs associated with evidence-based pathways in colon cancer.

This article was published as part of a special joint issue and also appears in the Journal of Oncology Practice.

Objective:

The goal of this study was to use 2 separate databases to evaluate the clinical outcomes and the economic impact of adherence to Level I Pathways, an evidence-based oncology treatment program in the treatment of colon cancer.

Patients and Methods:

The first study used clinical records from an electronic health record (EHR) database to evaluate survival according to pathway status in patients with colon cancer. Disease-free survival in patients receiving adjuvant treatment and overall survival in patients receiving first-line therapy for metastatic disease was calculated. The second study used claims data from a national administrative claims database to examine direct medical costs and use, including the cost of chemotherapy and of chemotherapy-related hospitalizations according to pathway status.

Results:

Overall costs from the national claims database—including total cost per case and chemotherapy costs—were lower for patients treated according to Level I Pathways (on- Pathway) compared with patients not treated according to Level I Pathways. Use of pathways was also associated with a shorter duration of therapy and lower rate of chemotherapy-related hospital admissions. Survival for patients on- Pathways in the EHR database was comparable with that in the published literature.

Conclusion:

Results from 2 distinct databases suggest that treatment of patients with colon cancer on-Pathways costs less; use of these pathways demonstrates clinical outcomes consistent with published evidence.

(Am J Manag Care. 2011;17(5 Spec No.):SP45-SP52)

This report assesses the impact of adherence to evidence-based treatment pathways on clinical outcomes, treatment complications, and cost of care in colon cancer.

  • This information can be used in practice and policy decisions when evaluating programs to address quality and value in cancer care, including the different data sources and the strengths and limitations of clinical and claims information.

Colorectal cancer (CRC) is the fourth most commonly diagnosed cancer in the United States and the second leading cause of cancer-related deaths.1 Survival has improved as a result of advancements in medical therapy and the availability of new treatments.2 The introduction of new agents has resulted in increased variability in treatment options and increased costs.3

Cancer treatment guidelines,4-6 or clinical pathways, have been developed to standardize treatment and improve quality of care. Little information exists, however, regarding the impact of adherence to pathways on clinical outcomes, treatment complications, and cost of care. Data suggest that the use of clinical pathways in oncology can produce improvements in some areas—including length of hospital stay, complications, and financial outcomes7—and are cost-effective in treating non-smallcell lung cancer.8

Level I Pathways is a set of treatment guidelines established as the foundation of an evidence-based oncology treatment program developed and led by physicians in the US Oncology (The Woodlands, TX) network. It uses only high-level evidence to assess efficacy, toxicity, and cost when recommending therapies. Treatments are updated regularly by a multidisciplinary task force in collaboration with disease research committees and practicing community oncologists. Recommendations are based on lines of therapy and are integrated into the iKnowMed (iKM) electronic health records (EHR) system to provide point-of-care decision support.

The MedStat MarketScan database (Thomson Reuters, New York, NY) represents the inpatient and outpatient healthcare service use of individuals nationwide covered by the benefit plans of large employers, health plans, and government and public organizations. The MarketScan database links paid claims and encounter data to detailed patient information across sites and types of providers over time. The annual medical database includes private sector health data from approximately 100 payers.

Our objective was to conduct 2 separate studies by using the 2 databases to evaluate the clinical outcomes and the economic impact of adherence to Level I Pathways in colon cancer treatment. The first study used clinical records from the iKM EHR database to evaluate survival. To address concern that pathways use may adversely impact quality of care in this study, we examined survival according to pathway status. In the second study, we applied a set of pathway rules to a separate cohort of patients with colon cancer from the MedStat national claims database to examine the cost of chemotherapy and chemotherapy-related adverse events.

Patients and Methods

Patient Identification and Characterization From the EHR Database

The first study was a retrospective cohort design identifying iKM EHR patients with a primary colon cancer diagnosis and initiating an adjuvant line of therapy or a first-line chemotherapy regimen for metastatic disease between July 1, 2006 and June 31, 2007 at US Oncology network practices. Patients who were in the middle of treatment, who were starting with second-line therapy or beyond, or whose regimens were unassessable (as a result of missing or conflicting information) for pathway status were excluded. Using clinical data from the EHR and Pathways reporting center, chemotherapy regimens were electronically assigned a pathway status. Patients were classified as on-Pathway if all regimens during the study period were consistent with the Level I Pathways recommendations for colon cancer. Patients were classified as off- Pathway if treatments were not consistent with pathways or if treatments changed from on- to off-Pathway or vice versa.

Patient Identification and Characterization From a National Claims Database

The second study used MedStat 2005, 2006, and 2007. A retrospective cohort design was used to query this large, commercial insurance database containing data for approximately 4.9 million insured lives with patient characteristics of 24 months of comprehensive fee-for-service coverage (medical and prescription coverage but no capitation) in 2005 and 2006 and at least 1 day of coverage in 2007, younger than age 70, and an active employee or spouse of an active employee. All patients with colon cancer who initiated cytotoxic chemotherapy during the first 6 months of 2006 and had not received chemotherapy in the 12 months before their initial chemotherapy date were identified. Colon cancer was identified by the presence of International Classification of Disase, Ninth Revision (ICD-9) 153.xx in 2006 with 2 or more examination and management physician claims or 1 or more inpatient facility or emergency department facility claims. Treatment was classified as adjuvant if patients had colon resection followed by chemotherapy that began within 60 days after resection; all other patients were considered metastatic. For each patient, the data were organized to produce a timeline of care, including details by dates of surgery, radiation therapy, and chemotherapy. Chemotherapy was categorized into regimens and lines of therapy. The care was compared manually with the 2006 Level I Pathways for colon cancer. Pathway status was calculated on the basis of cycle 1 of each line of therapy using drug combination rules. For example, if a Pathways regimen consisted of 3 drugs (eg, A B C), all had to be present for the patient to be considered on-Pathway. If a drug was omitted (eg, A C only), added (eg, A B C D), or substituted (eg, A B X), the regimen was considered off-Pathway. A nurse and team of pharmacists experienced in clinical oncology and pathways reviewed the timelines. Patients who received any care off- Pathway were classified as being off-Pathway.

Survival Analysis

Disease-free survival (DFS) and overall survival (OS) were compared between patients treated on-Pathway and off-Pathway by using the Kaplan-Meier method and an intentionto- treat analysis. DFS was calculated in patients starting an adjuvant line of therapy from chemotherapy initiation date to recurrence date, death, or censored for last date of contact. Because of the limited number of patients with stage II disease and pathway treatment changes occurring during the study period, these patients were excluded from the DFS analysis given that the groups could not be reliably separated according to treatment regimen. For patients starting firstline therapy for metastatic disease, OS was calculated from the chemotherapy initiation date to the date of death or last contact. Date of initiation of chemotherapy was defined as the date the chemotherapy regimen was ordered. Because of difficulty in obtaining radiology reports, the date of initiation of subsequent chemotherapy after adjuvant therapy was used to indicate relapse. Date of death, if available, was obtained from the EHR. If no relapse or death was noted, the last date of contact with documented patient vital signs was used.

Cost and Use Analysis

Allowed amounts (before cost sharing) from MedStat were tabulated during the chemotherapy period, which was from the date of the first chemotherapy to the last chemotherapy plus 30 days or until December 31, 2007, whichever came first, with a maximum of 18 months of total observation. Total patient care costs and chemotherapy costs, including oral and infused products, were tabulated. In addition, chemotherapyrelated hospital admissions during the chemotherapy period were tabulated. These were defined by the presence on the claim of a likely chemotherapy-related adverse effect as the primary diagnosis or the primary designation of a cancer diagnosis with 1 of the secondary diagnoses as a chemotherapy-related adverse effect. Statistical significance was calculated by using the Wilcoxon rank sum test.

Results

Clinical Results From an EHR Database

Tables 1

2

The study included 910 patients from 11 states with a diagnosis of colon cancer who met the EHR eligibility criteria. During the study period, 433 patients (48%) initiated djuvant therapy, and 477 (52%) initiated first-line therapy for metastatic disease. Of the total study population, 756 patients (83%) were treated on-Pathway, and 154 (17%) were treated off-Pathway. All patients had at least 35 months of observation with data through May 31, 2010. Patient characteristics of the adjuvant and metastatic populations according to pathway status are listed in and , respectively.

DFS: Adjuvant Chemotherapy

Figure 1

DFS was calculated for the 338 patients with stage III disease initiating adjuvant therapy during the study period. Eighty-five events (recurrence or death) occurred. Approximately 23% of on-Pathway patients (71 of 313) and 56% of off-Pathway patients (14 of 25) experienced an event. Median DFS was 26.9 months for off-Pathway patients and has not yet been reached for on-Pathway patients (; P <.05; hazard ratio, 4.98; 95% confidence interval [CI], 2.11 to 11.74). The estimated 1-year, 2-year, and 3-year DFS for on-Pathway and off-Pathway patients was 91% versus 72%, 80% versus 51%, and 73% versus 41%, respectively.

OS: Metastatic Disease

Figure 2

OS was calculated for the 477 patients initiating first-line therapy for metastatic disease. A total of 229 deaths occurred. Approximately 47% of on-Pathway patients (194 of 412) and 54% of off-Pathway patients (35 of 65) died. Median OS for on-Pathway was 26.9 versus 20.1 months for off-Pathway (; P = .03; hazard ratio, 1.57; 95% CI, 1.04 to 2.39). The 1-year estimated survival for on-Pathway was 80%; it was 74% for off-Pathway.

Economic Analysis: Cost and Use

Table 3

Table 4

From the MedStat national claims database, 220 patients with colon cancer treated with chemotherapy were identified who met study criteria. lists patient demographics. Classification of patients receiving adjuvant therapy or as having metastatic disease and as receiving treatment that adhered to Level I Pathways for colon cancer was determined by examining details of therapy as represented in the claims. This examination showed that 41% of patients were treated on-Pathway and 59% were treated off-Pathway. Cost and use results are listed in . For adjuvant treatment, total costs, chemotherapy costs, and the chemotherapy period were significantly lower for patients on-Pathway (P <.05). Costs per case and per patient per month were lower for patients on-Pathway, which was statistically significant for patients on adjuvant therapy. The chemotherapy-related admissions werelower in both adjuvant and metastatic on-Pathway patients, but this did not achieve statistical significance.

Discussion

Little is known regarding the extent of adoption or the impact of cancer treatment guideline adherence on survival, toxicities, and costs. We studied the impact of adherence to Level I Pathways for colon cancer on survival by using an EHR database (iKM) with almost 3 years of follow-up for all patients. Pathways are designed for common types of cancer rather than rare or complex cases. Our analyses revealed that Level I Pathways provide efficacious treatment options with benefits in terms of costs. Narrowing treatment options did not compromise care or negatively impact survival for on-Pathway patients. Additionally, outcome results were comparable with the published literature, which support our findings.2,9,10

We also assessed the direct medical costs associated with adherence to Level I Pathways for colon cancer by using national claims data from multiple payers. The retrospective claims analysis shows that costs are lower for patients on- Pathway. Use of pathways is also associated with a shorter duration of therapy and lower rate of chemotherapy-related admissions. The lack of statistical significance for costs in the metastatic cancer results was not surprising given the small number of on-Pathway patients. Notably, adherence to pathways was 41% from the national claims database compared with 83% from the EHR database. This may be an effect of integration of the evidence-based treatment recommendations directly into the EHR. Additionally, exclusion of unassessable regimens can increase the adherence rate if many of those excluded were off-Pathway. In the absence of direct methodology validation, each retrospective analysis is an important supplement to the other and is consistent with an earlier publication of clinical data related to Level I Pathways that reported lower chemotherapy costs for on-Pathway versus off-Pathway treatment in non-small-cell lung cancer.8 This study provides additional support for the concept that Level I Pathways can lower costs and adhere to high-level evidence.

Since 1996, 6 new therapeutic agents have been approved for colon cancer, which has resulted in increased treatment options, variability, and costs.11 Three new agents are monoclonal antibodies and contribute upwards of a 340-fold increase in the cost of chemotherapy compared with previous treatments.11,12 This information has quality and financial implications for patients and the healthcare system.13 Of note, the Level I Pathways include these newer chemotherapy agents when the evidence supports their use.

Different approaches can be used to assess adherence to cancer treatment guidelines. EHR and claims analysis each have strengths and limitations. Payer databases can capture diagnoses, medical service procedures, drug therapy use, and costs, including inpatient, outpatient, and prescription drug costs. This makes it possible to capture all patient treatments and medical costs. Large payer databases may provide significant sample sizes even after applying stringent conditions and may provide findings that are not specific to certain practice sites or regions. However, there are inherent restrictions in the analyses of claims data, regardless of whether an analysis is prospective or retrospective. Claims analysis suffers from the absence of important clinical information including cancer stage and line of therapy, which makes assessment of adherence challenging. Claims data may also contain coding errors, inconsistencies, or omissions. Patient outcomes, including mortality, are difficult to determine. Commercial insurance data make it possible to identify when an individual loses coverage, but it is generally not possible to determine whether that loss results from death, switches in insurance coverage, or another cause. Conversely, patient charts and EHRs can capture clinical information that claims cannot, such as staging, line of therapy, histologic subtypes, and tumor marker information. In an era of personalized medicine, these data are important when assessing adherence to guidelines. Importantly, clinical outcomes and mortality information may also be obtained. However, EHRs may not provide information regarding inpatient admissions, services obtained outside the practice, or outpatient pharmacy use.

Limitations of these results should be noted. This report comprised 2 independent studies with distinct patient populations; therefore, direct comparisons of results between the studies cannot be made. The reasons for treating off-Pathway are not known. Given that risk factors in these patients are not completely understood and numbers are limited until databases are expanded, survival may not be the best outcome measurement for off-Pathway patients on the basis of EHR databases. This is an area for future exploration. The EHR population included Medicare patients, and the claims population did not. Rules were applied to claims to determine pathway status. Surgery, in a 60-day window, is an arbitrary indicator of adjuvant therapy and may include patients with metastatic disease. Pathway status was assigned electronically in the EHR compared with manually in the claims analysis. In addition, on the basis of EHR confines, this was an intention-to-treat analysis using treatment initiation dates that were based on when regimens were ordered for patients. Although Level I Pathways includes chemotherapy choices to address comorbidities, a selection bias for the off-Pathway cohort in either study cannot be excluded. A randomized trial of on- versus off-Pathway treatment is unlikely, so we are left with retrospective evaluations. The strength of these separate analyses is in the ability to query comprehensive EHR and national claims data to separately evaluate the impact of adherence to Level I Pathways on clinical outcomes, medical costs, and medical use. Additional research in these areas will aid in future comparative effectiveness studies

Keys to successful guidelines as discussed by Smith et al7 include inclusion of clinicians in the development with the opportunity for critique during the development process, dissemination through targeted communication and educational efforts, implementation through patient-specific reminders, and accountability for process and outcomes.7 These components all exist in the Level I Pathways program.

In conclusion, 2 independent implications can be drawn from these analyses: treatment on-Pathway costs less, and no decrease in patient survival was observed. This study represents the exploration of comparative effectiveness by using novel methods of mining retrospective databases to make assessments about care delivery and value. Innovative collaborations between providers and payers to share clinical and economic information will help address quality and cost issues in the treatment of cancer.

Acknowledgment

We thank Lina Asmar, PhD, and Ann Morcos, MA, ELS, for their biostatistics and medical writing assistance; Sheetal Sheth, PharmD, and Michael Forsyth, RPh, for assistance with data analysis; Roger Anderson, DrPH, the US Oncology Pathways Task Force, and the Clinical Content Development Team for their assistance with study materials and methods.

Authors’ Disclosures of Potential Conflicts of Interest. Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: J. Russell Hoverman, US Oncology (C); Debra A. Patt, US Oncology (C); Michael Kolodziej, US Oncology (C); Marcus A. Neubauer, US Oncology (C); Roy A. Beveridge, US Oncology (C).

Consultant or Advisory Role: Thomas H. Cartwright, Genentech (C), Eli Lilly (C), Amgen (C), sanofi-aventis (C); Kathryn Fitch, US Oncology (C); Bruce Pyenson, US Oncology (C).

Stock Ownership: None.

Honoraria: Thomas H. Cartwright, Roche, Amgen, Eli Lilly, sanofi-aventis; Roy A. Beveridge, US Oncology.

Research Funding: Roy A. Beveridge, US Oncology.

Expert Testimony: None.

Other Remuneration: None.

Author Contributions

Conception and design: J. Russell Hoverman, Thomas H. Cartwright, Janet L. Espirito, Terrance J. Kopp, Michael Kolodziej, Kathryn Fitch, Bruce Pyenson, Roy A. Beveridge.

Administrative support: Debra A. Patt, Janet L. Espirito, Jody S. Garey, Terrance J. Kopp, Roy A. Beveridge.

Provision of study material or patients: Thomas H. Cartwright, Jody S. Garey.

Collection and assembly of data: Debra A. Patt, Janet L. Espirito, Matthew P. Clayton, Jody S. Garey.

Data analysis and interpretation: J. Russell Hoverman, Thomas H. Cartwright, Debra A. Patt, Janet L. Espirito, Michael Kolodziej, Kathryn Fitch, Bruce Pyenson, Roy A. Beveridge.

Manuscript writing: J. Russell Hoverman, Thomas H. Cartwright, Debra A. Patt, Janet L. Espirito, Michael Kolodziej, Kathryn Fitch, Bruce Pyenson.

Final approval of manuscript: J. Russell Hoverman, Thomas H. Cartwright, Debra A. Patt, Janet L. Espirito, Matthew P. Clayton, Jody S. Garey, Terrance J. Kopp, Michael Kolodziej, Marcus A. Neubauer, Kathryn Fitch, Bruce Pyenson, Roy A. Beveridge.

Address Correspondence to: J. Russell Hoverman, MD, PhD, Texas Oncology, Suite 500, 12221 Merit Dr, Dallas, TX 75251; e-mail: russell.hoverman@usoncology.com.

1. Jemal A, Siegel R, Xu J, et al: Cancer statistics, 2010. CA Cancer J Clin. 60:277-300, 2010.

2. Kopetz S, Chang GJ, Overman MJ, et al: Improved survival in metastatic colorectal cancer is associated with adoption of hepatic resection and improved chemotherapy. J Clin Oncol. 27:3677-3683, 2009.

3. Meropol NJ, Schulman KA: Cost of cancer care: issues and implications. J Clin Oncol. 25:180-186, 2007.

4. National Comprehensive Cancer Network: NCCN Clinical Practice Guidelines in Oncology: colon cancer. http://www.nccn.org/professionals/physician_gl/f_guidelines.asp.

5. Labianca R, Nordlinger B, Beretta GD, et al: Primary colon cancer: ESMO clinical practice guidelines for diagnosis, adjuvant treatment and follow-up. Ann Oncol. 21:v70-v77, 2010 (suppl 5).

6. Desch CE, Benson AB 3rd, Somerfield MR, et al: Colorectal cancer surveillance: 2005 update of an American Society of Clinical Oncology practice guideline. J Clin Oncol. 23:8512-8519, 2005.

7. Smith TJ, Hillner BE: Ensuring quality cancer care by the use of clinical practice guidelines and critical pathways. J Clin Oncol. 19:2886- 2897, 2001.

8. Neubauer MA, Hoverman JR, Kolodziej M, et al: Cost effectiveness of evidence-based treatment guidelines for the treatment of nonsmall- cell lung cancer in the community setting. J Oncol Pract. 6:12-18, 2010.

9. André T, Boni C, Mounedji-Boudiaf L, et al: Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. N Engl J Med. 350:2343-2351, 2004.

10. Sargent DJ, Wieand HS, Haller DG, et al: Disease-free survival versus overall survival as a primary end point for adjuvant colon cancer studies: individual patient data from 20,898 patients on 18 randomized trials. J Clin Oncol. 23:8664-8670, 2005.

11. Ferro SA, Myer BS, Wolff DA, et al: Variation in the cost of medicationsfor the treatment of colorectal cancer. Am J Manag Care. 14:717-725, 2008.

12. Schrag D: The price tag on progress: chemotherapy for colorectal cancer. N Engl J Med. 351:317-319, 2004.

13. Howard DH, Kauh J, Lipscomb J: The value of new chemotherapeutic agents for metastatic colorectal cancer. Arch Intern Med. 170:537- 542, 2010.