Health Plan Retention and Pharmacy Costs of Newly Diagnosed Patients with Chronic Kidney Disease in a Managed Care Population

Chronic kidney disease (CKD) affects approximately 26 million people in the United States.¹ Diabetes and hypertension cause up to two thirds of all new CKD cases.^1,2 According to Medicare policy, health plans are financially responsible for the care of CKD patients for up to 33 months after they have reached the final stage of end-stage renal disease (ESRD).³ Data from the Institute for Health and Productivity Management 2001 database show that treatment costs nearly double from one stage of CKD to the next.⁴ The stages of CKD are defined based on the glomerular filtration rate (GFR) as determined by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative guidelines.⁵

Insurance claims for patients with CKD average $5000 to $12,000 per patient per year (PPPY) as a patient proceeds from stage 1 (GFR =90 mL/min/1.73 m²) to stage 2 (GFR 60-89 mL/min/1.73 m²), $15,000 to $28,000 PPPY from stage 3 (GFR 30-59 mL/min/1.73 m²) to stage 4 (GFR 15-29 mL/min/1.73 m²), and exceed $70,000 PPPY once the patient reaches stage 5 (GFR <15 mL/min/1.73 m²).⁴ Progressive CKD presents a burden to employers as suggested by researchers who examined ESRD-related nonmedical costs for employers.⁶ It was estimated that employers may incur life insurance costs of $55,055 per ESRD-related death, disability insurance costs of $31,617 per ESRD disability, and worker replacement costs of $27,869 per ESRD-related lost worker.⁶

From a managed care perspective, it has been suggested that controlling and stabilizing the main comorbidities of CKD—diabetes and hypertension—may slow the progression of CKD, which would result in a reduction of healthcare costs.⁷ Factors that may be associated with optimal quality of care for patients with CKD during the 12-month predialysis phase have been identified, including predialysis erythropoietin therapy, nephrology referrals, and phosphate binder/vitamin D administration. As many as 48.7% of patients did not have any interventions associated with optimal care, suggesting a need for proactive management of CKD in the managed care setting to reduce utilization, while improving patient outcomes.⁸ Similarly, a retrospective claims analysis of 27 health plans in 19 states evaluated resource utilization in 3 defined time periods9:

• Predialysis: months 2 to 6 before initial dialysis
• Peridialysis: 30 days before and 30 days after dialysis
• Postdialysis: months 2 and 3 after initial dialysis.

Per patient per month charges were highest in the peridialysis period ($35,292 vs $4265 for predialysis and $15,399 for postdialysis), and treatments with nutritional supplements and medications such as angiotensin-converting enzyme inhibitors and erythropoietin were suboptimal.⁹ A directive has been issued for managed care plans to manage CKD through early intervention to improve outcomes and reduce costs.^10-12 Although data are available that establish the cost impact of CKD, as well as identify gaps in the treatment of CKD patients at all stages, data to characterize retention among newly diagnosed CKD patients in managed care plans are limited.

Information regarding retention of CKD patients will enable managed care plans to understand the potential impact of treating predialysis CKD as a chronic illness on the plan and the impact on its members. Such data may encourage plans to implement early-intervention strategies and potentially minimize costly expenditures in later stages.

The purpose of this study was to examine the health plan retention rates and pharmacy costs among newly diagnosed CKD patients, including those with diabetes or hypertension, from a managed care perspective. The main focus was to look at the length of time patients with CKD, CKD and diabetes, and CKD and hypertension remain within a health plan after the initial CKD diagnosis, and compare the direct costs of CKD alone and the costs of CKD plus these closely related comorbidities.

Methods
Data Source
De-identified medical and pharmacy claims from the PharMetrics patient-centric database¹³ between January 1995 and December 2006 were used to conduct this analysis. The PharMetrics database was selected because it is largely representative of the commercially insured population in the United States. This longitudinal database captures de-identified medical and pharmacy claims from 85 health plans that are submitting complete patient-level data for more than 45 million managed care lives. The average member enrollment across the entire database is 2 years. Data elements used in this analysis include health plan product type, health plan payer type, enrollment records, patient demographics, diagnoses, Episode Treatment Groups (ETG) codes, and prescription records.

Study Design

Figure 1

We utilized a retrospective cohort consisting of 3 subgroups of newly diagnosed patients with CKD: (1) CKD only, (2) CKD and diabetes, and (3) CKD and hypertension. All patients were at least 18 years old as of the date of their initial CKD diagnosis, had initiated health plan enrollment between January 1, 1995, and September 2001, and had at least 1 CKD claim, using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes: 585.XX, 592.XX, 593.9, including a 180-day preindex period with no evidence of CKD (Figure 1).

Patients were excluded if they had missing member data or enrollment data of less than 6 months. Patients identified after 2001 were not included, because the objective was to have a 5-year follow-up period for the identified diagnosis. Each patient was followed for 5 years from index diagnosis, and the length of enrollment was captured at the patient level.

Subgroups were identified by segmenting the patients with CKD according to evidence of the selected comorbidities—diabetes and hypertension. Diabetes and hypertension were identified by the presence of corresponding ETG codes. The ETG coding method is an illness classification system that identifies and classifies an entire episode of care for a patient, taking into account all medical care delivered in the inpatient and outpatient setting collectively for that episode.¹⁴ The ETG system considers patient age, comorbidities, procedures, clinical complications, and pharmaceutical claims in a synergistic manner to group patients by total clinical activity.

ETG coding was selected for this study over ICD-9-CM codes, because ETG codes more accurately and completely represent the patient episodic healthcare profile. ETG codes allow researchers to capture underdiagnosed cases that would otherwise be missed by ICD-9-CM codes. ETG codes are already formatted and readily retrievable from the PharMetrics patient-centric database.¹³ Patients with no evidence of diabetes or hypertension were categorized as the CKD-only group. CKD patients with ETG codes 278 to 281 and 908-4 were categorized as the CKD and diabetes group. CKD patients with ETG codes 27 to 30 were categorized as the CKD and hypertension group.

Outcomes Measures
The overall duration of health plan enrollment was reported in days and months, as a continuous variable, for each subgroup based on patient level enrollment records. Enrollment status for each patient was analyzed and aggregated at annual intervals, and reported as a yearly categorical variable for each subgroup.

This study looked at direct healthcare costs and pharmacy costs, which together comprise total medical costs. Annualized direct healthcare costs were calculated based on health plan payments for inpatient, emergency department visits, home health visits, office visits, laboratory testing, and other outpatient claims. Costs were divided into total inpatient, total outpatient, and total medical costs for each subgroup. The annualized cost of total pharmaceutical treatment was calculated based on health plan payments for all prescription medications. Costs were summarized for each subgroup and presented in US dollars.

Statistical Analysis
Descriptive statistics were provided for both continuous and categorical variables by subgroup and overall. All statistical analyses were performed using SAS release 9.1.3 (SAS Institute, Cary, NC).

Results
Study Population
A total of 31,917 eligible patients were identified and comprised the newly diagnosed CKD cohort. The study population consisted of 8836 CKD-only patients, 11,252 patients with CKD and diabetes, and 20,836 patients with CKD and hypertension. The CKD and diabetes and CKD and hypertension subgroups were not mutually exclusive, and had overlapping patients. Among 23,081 patients identified with comorbidities, 9007 (39%) patients had both diabetes and hypertension.
 

Table 1

Table 1 lists the study population baseline characteristics for the subgroups. The average age for all eligible patients with CKD was 55 years; of these patients, 51.2% were men. Patients with CKD and diabetes (average, 58.3 years) and those with CKD and hypertension (average, 59.2 years) were older than the CKD-only (average, 45.6 years) patients. The proportion of men was higher in the CKD and diabetes (53.8%) and CKD and hypertension (52.6%) subgroups compared with the CKD-only subset (47.2%).

Slight differences were seen in the initial diagnosis of CKD by physician specialty. There was a trend toward a greater proportion of patients with CKD and diabetes and patients with CKD and hypertension diagnosed with CKD by a nephrologist compared with the CKD-only subgroup. A large number of patients in each subgroup were initially diagnosed with CKD by a physician who is not specialized in cardiology, nephrology, endocrinology, or internal medicine/general family practice. The top identifiable specialties included in the "other" category were urology and emergency department physicians; however, the majority of physician coding was not specified.

Geographic distribution reflected representation from all 4 regions of the United States, with most patients residing in the Midwest. The majority of the patients in each subgroup belonged to a health maintenance organization.
 

Figure 2

Figure 2 outlines the distribution of patients by payer type. A greater proportion of CKD-only patients (63%) was enrolled in a commercial plan compared with those with CKD and diabetes (46%) or with CKD and hypertension (44%). Conversely, a greater proportion of patients with CKD and diabetes (40%) and CKD and hypertension (40%) had coverage through a Medicare Risk plan compared with CKD-only patients (13%).

Duration of Enrollment

Table 2

The average duration of enrollment was slightly longer for the CKD and diabetes (1163 days; 39 months) and the CKD and hypertension (1190 days; 40 months) subgroups compared with the CKD-only (1079 days; 36 months) subgroup. Overall, newly-diagnosed CKD patients remained enrolled in their health plan at 3 years postdiagnosis. Table 2 summarizes the enrollment status at yearly intervals for the subgroups and the overall group.

When grouping the patients into subgroups of CKD only, CKD and hypertension, or CKD and diabetes, we looked at their entire diagnostic phase and follow-up period. For example, for patients to be classified as CKD only, they could not have hypertension or diabetes mellitus anytime on or after their first diagnosis of CKD up to 5 years of follow-up period.

Costs

Table 3

Figure 3

Annual direct healthcare costs are outlined in Table 3 and Figure 3 . The mean annual direct healthcare costs were substantially greater for patients with CKD and diabetes ($20,165) and those with CKD and hypertension ($17,612) compared with patients with CKD only ($9390).
 

Figure 4

Pharmacy. The mean annual pharmacy costs (Table 3, Figure 4) were substantially higher for the patients with CKD and diabetes ($2409) and CKD and hypertension ($2185) compared with the CKD-only ($904) patients.

Medical. The mean annual total medical costs (Table 3) were substantially greater for patients with CKD and diabetes ($22,444) or CKD and hypertension ($19,667) compared with patients with CKD only ($10,170).

Discussion
The study used ETG codes to classify the patients into 3 groups; those with CKD and hypertension and those with CKD and diabetes mellitus were given this designation at the beginning of the study and were then compared with the CKD-only group regarding enrollment duration and cost. This project was undertaken to identify the proportion of patients with newly diagnosed CKD that remain in the same health plan over time. An assumption was made that the longer a patient remains within a health plan, the better the opportunity to implement and assess the impact of early intervention strategies for disease management and overall quality of health.

Smith and colleagues documented that per-patient costs increased as the stages of CKD progressed.15 Our results of increased pharmacy costs for CKD patients with disease-related comorbidities agree with Smith and colleagues' findings, which demonstrated that costs for CKD patients were much higher in those with diseaserelated comorbidities compared with those with CKD alone. Delay of disease progression to the next stage could, therefore, also have financial benefits to health plans. Implementation of an aggressive disease-specific approach by a health plan could impact patient outcomes by a possible delay of disease progression as well.

In this study, we evaluated 3 groups of newly diagnosed CKD patients stratified based on the documented coexistence of hypertension or diabetes. Patients with CKD and comorbidities tended to stay with their same health plan for approximately 3 months longer than patients with CKD only. To ensure that CKD patients were accurately identified, a frequency distribution was run, evaluating the number of patients in each subgroup (ie, CKD only, CKD and diabetes, and CKD and hypertension) with 1 claim for CKD versus those with more than 1 claim. The subgroups were similar in that approximately 70% of patients in each subgroup had at least 2 claims for CKD, providing additional validity to the diagnosis of CKD.

Approximately 66% of CKD-only patients were enrolled in commercial plans, whereas a larger proportion of the patients with CKD and diabetes and CKD and hypertension were enrolled in Medicare Risk plans. The average retention time may be influenced or directed by the Medicare timeline law that dictates when Medicare coverage begins for a CKD dialysis patient. As the law extends this timeframe from the current 33 months to the proposed 42 months, retention time in health plans will most likely expand, further necessitating the need to intervene as early as possible in these at-risk employee populations.¹⁶

Early intervention in these populations is paramount to preventing or delaying disease progression and, in parallel, to potential cost reduction for the health plan, employers, and patients/members. Our findings, which have not been documented extensively in previous studies, support the notion that newly diagnosed patients with CKD and those with diseaserelated comorbidities retain their healthcare plans for considerable periods.

Limitations
This study design has several limitations. Our goal was to highlight the amount of time patients who are actually diagnosed with CKD remain in plans, which provide a strong incentive to treat it appropriately rather than assume that patients would not be in the plan long enough to reap the benefits of treatment. We acknowledge that the number of patients diagnosed with CKD in the real world is well below what would be found if proper screening was conducted. We also acknowledge that CKD was likely more prevalent than it appears in these data we evaluated, although patients with CKD who are not diagnosed may move between plans at similar rates as those without CKD.

In addition, there were potential overlaps between the CKD and diabetes and CKD and hypertension subgroups. Factors for patient disenrollment were not available due to the retrospective nature of the study. Only newly diagnosed CKD patients were included in the study; therefore, retention periods for existing CKD patients and a healthy non-CKD cohort were not evaluated. A diagnosis of CKD was captured during the identification period only (July 1, 1995-September 30, 2001), so patients with a new CKD diagnosis during the 5-year follow-up period were not evaluated.

The study design also did not account for patients whose diagnosis of CKD might have been ruled out by further investigation. The retention rates reported may not be representative of all CKD patients, because they may be different for patients who had a previously established CKD diagnosis (not newly diagnosed). Also, retention rates and costs were not evaluated within and between subgroups for differing stages of CKD and hypertension or severity of diabetes (eg, requiring oral or injectable medications). Direct comparison between groups in similar plan types was also not evaluated.

This study does not account for access to healthcare differences and its impact on costs among different plans. Finally, the use of ETG codes in this analysis limited the ability to compare it with other studies using traditional diagnosis and procedure costs.

Conclusion The majority of newly diagnosed CKD patients retain their health plan affiliation for a considerable period, including those with diabetes or hypertension. Presence of these comorbidities resulted in increased total direct healthcare costs. Results of this study further support the need for early intervention in CKD. Managed care plans may benefit from earlier management strategies for CKD, because these patients could contribute to increased healthcare utilization and cost over the long-term.

Acknowledgment
This study was funded by Centocor Ortho Biotech Services, LLC.

Disclosure Statement
All the authors are employed by Centocor Ortho Biotech Services, LLC.

References

1. National Kidney Foundation. The facts about chronic kidney disease (CKD). www.kidney.org/kidneydisease/ckd/index.cfm#facts. Accessed June 28, 2008.
2. The American Society of Nephrology. Facts and Statistics. www.asn-online.org/facts_and_statistics/faq.aspx. Accessed June 28, 2008.
3. Centers for Medicare & Medicaid Services. Medicare Coverage of Kidney Dialysis and Kidney Transplant Services. May 2008. www.medicare.gov/Publications/Pubs/pdf/10128.pdf. Accessed June 28, 2008.
4. Sullivan S. Employer challenges with the chronic kidney disease population. J Manag Care Pharm. 2007;13(9 suppl D):S19-S21.
5. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis. 2002;39(2 suppl 1):S1-S266.
6. Kamal-Bahl SJ, Pantely S, Pyenson B, Alexander CM. Employer-paid nonmedical costs for patients with diabetes and end-stage renal disease. Prev Chronic Disease. 2006;3:A83. Epub 2006 Jun 15.
7. Rasu RS, Crawford T, Manley HJ, Balkrishnan R. Treatment of hypertension and diabetes mellitus in patients with chronic kidney disease: a review. Expert Opin Pharmacother. 2007;8:2543-2551.
8. London R, Solis A, Goldberg GA, et al. Examination of resource use and clinical interventions associated with chronic kidney disease in a managed care population. J Manag Care Pharm. 2003;9:248-255.
9. Robbins JD, Kim JJ, Zdon G, et al. Resource use and patient care associated with chronic kidney disease in a managed care setting. J Manag Care Pharm. 2003;9:238-247.
10. Chronic kidney disease: stating the managed care case for early treatment. Discussion and consensus of presentations of economic analyses, managed care organization case studies, and opportunities for intervention in a managed care setting. May 3-5, 2001, Chicago, Illinois. Am J Manag Care. 2002;8(4 suppl):S112-S120.
11. Weiner DE. Causes and consequences of chronic kidney disease: implications for managed health care. J Manag Care Pharm. 2007;13(3 suppl):S1-S9.
12. Wright A. Preparing MCOs to manage chronic kidney disease. Manag Care. 2003;Spec No:13-6; discussion 17-20.
13. IMS Health. PharMetrics patient-centric database. Watertown, MA; 2008.
14. Ingenix. Symmetry episode treatment groups: measuring health care with meaningful episodes of care. 2007. www.ingenix.com/content/attachments/SymmetryETG_WhitePaper.pdf. Accessed June 27, 2008.
15. Smith DH, Gullion CM, Nichols G, et al. Cost of medical care for chronic kidney disease and comorbidity among enrollees in a large HMO population. J Am Soc Nephrol. 2004;15:1300-1306.
16. Geisel J. House panel approves renal care cost-shift. Business Insurance. July 30, 2007. www.businessinsurance.com/article/20070730/NEWS/200010770. Accessed October 16, 2008.