Economic Evaluation of a Disease Management Program for Chronic Obstructive Pulmonary Disease

Abstract

Background: The data on cost savings with disease management (DM) in chronic obstructive pulmonary disease (COPD) is limited. A multicomponent DM program in COPD has recently shown in a large randomized controlled trial to reduce hospitalizations and emergency department visits compared to usual care (UC). The objectives of this study were to determine the cost of implementing the DM program and its impact on healthcare resource utilization costs compared to UC in high-risk COPD patients. Materials and Methods: This study was a post-hoc economic analysis of a multicenter randomized, adjudicator-blinded, controlled, 1-year trial comparing DM and UC at 5 Midwest region Department of Veterans Affairs (VA) medical centers. Health-care costs (hospitalizations, ED visits, respiratory medications, and the cost of the DM intervention) were compared in the COPD DM intervention and UC groups. Results: The composite outcome for all hospitalizations or ED visits were 27% lower in the DM group (123.8 mean events per 100 patient-years) compared to the UC group (170.5 mean events per 100 patient-years) (rate ratio 0.73; 0.56–0.90; p < 0.003). The cost of the DM intervention was $241,620 or $650 per patient. The total mean ± SD per patient cost that included the cost of DM in the DM group was 4491 ± 4678 compared to $5084 ± 5060 representing a $593 per patient cost savings for the DM program. Conclusions: The DM intervention program in this study was unique for producing an average cost savings of $593 per patient after paying for the cost of DM intervention.

Keywords::

• Chronic Obstructive Pulmonary Disease
• Economics
• Education
• Disease Management

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is a major public health problem associated with substantial economic and societal costs. COPD currently affects 30 million Americans with an annual estimated cost of $37 billion (1, 2). Hospitalizations account for the largest portion of healthcare expenditures for COPD. Depending on the health system, hospitalizations accounted for 54% to 70% of the total direct healthcare costs of COPD (3, 4).

Disease management (DM) is a comprehensive strategy for improving overall health status and reducing healthcare costs in chronic conditions through integrated care programs (5). Although a substantial number of studies evaluating different types of DM programs have been conducted in patients with COPD, only a limited number were able to demonstrate an improvement in health outcomes (6–9). DM is less effective in COPD than it is for other chronic conditions such as heart failure, asthma, depression, and diabetes (10). DM programs demonstrated to be successful in COPD had to include two or more components of the chronic care model (11). In addition, successful COPD DM programs were resource intensive and included relatively small sample sizes (6–9). Additionally, some DM programs incorporating self management showed improved patient outcomes with reduced costs while others showed a potential for economic benefit with increased patient case load (12, 13). Identifying cost-effective multicomponent COPD DM programs still remains a challenge (14).

More recently, a large-scale multicomponent multicenter COPD DM program implemented within a regional Veterans Affairs (VA) medical system was found to significantly reduce COPD-related hospitalizations and emergency department visits by 41%.(15). Other potential benefits included significant reduction in all cause hospital admissions and emergency department visits and significant improvement in the quality of life. Additionally, this study was unique that the DM intervention was relatively simple and easy to implement in a large group of patients that could be applicable to the real world.

To be able to routinely adopt this DM program for use in high risk COPD patients, the relative cost of the intervention along with potential cost savings from reduction in health care utilization need to be evaluated. The present study is a post-hoc economic evaluation of the multicenter VA COPD DM intervention. The data collected from the prospective, randomized comparison of the VA DM intervention was used to evaluate the cost of the intervention compared to usual care (UC).

METHODS

Study Design

The methods and outcomes of the VA COPD DM study have been previously described in full and are briefly summarized below ([12]). The study was a randomized, controlled trial conducted at 5 Veterans Affairs medical centers (Minneapolis MN, Iowa City IA, Des Moines IA, Omaha NE, and Sioux Falls, SD). COPD patients at high risk for exacerbations were randomly assigned to either DM or UC. Inclusion criteria included a diagnosis of COPD and one or more of the following during the previous year: 1) hospital admission or emergency department visit for COPD, 2) chronic long-term home oxygen use, or 3) a course of systemic corticosteroids for COPD.

Exclusion criteria included inability to have access to a home telephone line or sign a consent form, or any condition that would preclude effective participation in the study or likely to reduce life expectancy to less than a year. The primary study outcome was the combined number of hospitalizations or emergency department visits for COPD incurred by each patient during the 12-month follow-up period. Quality of life, as judged by self-reported respiratory health status using the St. George's Respiratory Questionnaire, was evaluated as a secondary outcome in the study.

Patients assigned to UC received a one-page handout containing a summary of the principles of COPD care according to published guidelines and the telephone number for the 24-h VA nursing helpline, a service available to all VA patients. Patients assigned to the DM arm attended a single 1- to 1.5-h group education session conducted by a respiratory therapist case manager. The case managers were experienced registered respiratory therapists who received a 1-day training session on the study protocol. The patient education session used educational material developed by the American College of Chest Physicians, that included general information about COPD, direct observation of inhaler techniques, a review and adjustment of outpatient COPD medications, smoking cessation counseling, influenza and pneumococcal vaccinations, encouragement of regular exercise, and instruction in hand hygiene. Formal cardiopulmonary rehabilitation was not available at any study site and was not offered as part of the study protocol.

Each DM subject received an individualized written action plan that included: (1) a description of the signs and symptoms of an exacerbation that should prompt initiation of self-treatment, (2) refillable prescriptions for prednisone and an oral antibiotic, (3) contact information for a case manager, and (4) the telephone number of the 24-h VA nursing helpline. Subjects were instructed to begin action plan medications that included antibiotics and steroids for symptoms that were substantially worse than usual.

The case manager made monthly phone calls to patients assigned to DM to reinforce the general principles of COPD management and the use of action plan. Patients were encouraged to call the case manager during regular working hours if they took action plan medications or if they had questions relating to their medical care. There were no regularly scheduled clinic visits for the remainder of the 1-year follow-up period and there was no tracking for the regular visits to the physicians.

Economic analysis

The cost analysis was conducted from the perspective of the VA healthcare system. There were two primary economic objectives of the analysis. The first was to determine the cost of the COPD DM program. The second objective of the economic analysis was to compare the direct healthcare costs during the 1-year follow-up in the 2 treatment groups. If the cost of the DM intervention plus the costs of the healthcare resources utilized was greater in the DM group than in the UC group, an incremental cost-effectiveness analysis was to be performed. The incremental cost-effectiveness analysis was designed to calculate the cost to avoid one all-cause hospitalization.

Costs included in the analysis for both groups were the direct medical care costs of hospitalizations, emergency department (ED) visits, and prescriptions for pulmonary drugs (steroids, antibiotics, bronchodilators). The cost of the DM intervention was applied only to patients in the DM treatment group. There were no regularly scheduled outpatient clinic visits in the randomized trial and costs of routine clinic visits were not assessed. Indirect healthcare costs were also not captured in the randomized trial. Costs were not discounted as the duration of the trial was ≤ 1 year.

Hospitalization costs were based on the VA Health Economics Resource Center (HERC) Average Cost Datasets for medical/surgical discharges from VA hospitals (13). Costs were based on the specific adjudicated DRG assigned for each hospital admission that occurred during the prospective study. Costs of ED visits were based on the actual costs of those visits derived from the VA Decision Support System (DSS) National Data Extract files. Costs of pulmonary drugs were also derived from the VA DSS pharmacy data set for outpatient pharmacy cost data. The cost of the DM intervention included the actual VA salary and fringe benefits of the respiratory therapist case managers (average annual salary $41,846 and 30% fringe benefits of $17,934 for a total salary and fringe of $59,780) and the direct costs of the printed educational material and the use of a telephone and telephone line at each site. There was no tracking for the actual time spent by the case managers with the patients.

To investigate the impact of assumptions made during the analysis and to test the robustness of results given cost and workload variation in the data input, two sensitivity analyses were performed. The case work load was varied for the respiratory therapist case managers. The respiratory therapists in this study managed an average of 93 patients (range 75–125) over the 12 months of the study. In a workload sensitivity analysis, the cost of the case manager salaries was adjusted based on a work load increased to include 125 and 150 patients per case manager over the course of the 12 months effectively reducing the required salaries to implement the DM program. We believe that managing 150 patients per case manager is realistic since this is the number of patients that are currently being managed by the case managers after the study ended.

A second sensitivity analysis replaced respiratory therapists as case managers with registered nurses. Registered nurses have higher salaries (annual salary $54,813 plus 30% fringe benefits of $23,491 for a total salary and fringe of $78,304), which may impact the cost effectiveness of the DM intervention. This analysis was introduced as many physician offices do not employ respiratory therapists, but are typically staffed by registered nurses.

Statistical analysis was performed using STATA (release 11, StataCorp LP, College Station, TX). The mean ± SD per patient costs for hospitalizations, ED visits, and pulmonary drug prescriptions was calculated for patients in each treatment group. Because the cost data was not normally distributed, the bootstrap method was used to estimate 95% confidence intervals to test for statistically significant differences in total cost between the treatment groups. To estimate these confidence intervals, the study population was resampled 1000 times. All statistical tests were two-sided with a significance level at 0.05.

RESULTS

The clinical results of the VA COPD DM study have been previously described in full, and are briefly summarized next (12). Then, 743 patients met the inclusion criteria with 371 randomized to UC and 372 randomized to DM. All patients were followed for 12 months or until their time of death if it occurred prior to the 12-month follow-up. The average duration of follow-up was 341 days for UC patients and 349 days for DM patients. The baseline characteristics of the 2 treatment groups were well matched with no significant differences between the groups.

At the end of the 1-year follow-up, the composite outcome for all hospitalizations or ED visits were 27% lower in the DM group (123.8 mean events per 100 patient-years) compared to the UC group (170.5 mean events per 100 patient-years) (rate ratio 0.73; 0.58–0.90; p < 0.003) (Table 1).

Table 1.  Hospitalizations and emergency department visits during follow-up

	Disease Management (n = 372)	Usual care (N = 371)	Rate ratio (95% CI)	P-value
COPD
Hospitalization or ED visit	48.4	82.2	0.59 (0.44, 0.78)	< 0.001
Hospitalization	27.6	39.8	0.69 (0.47, 1.01)	0.08
ED visit	20.8	42.4	0.49 (0.33, 0.72)	< 0.001
Cardiac or non-COPD pulmonary
Hospitalization or ED visit	18.3	26.8	0.68 (0.44, 1.05)	0.06
Hospitalization	8.7	17.0	0.51 (0.29, 0.90)	0.01
ED visit	9.6	9.8	0.98 (0.56, 1.70)	0.93
Other than pulmonary or cardiac
Hospitalization or ED visit	56.6	60.9	0.93 (0.69, 1.25)	0.58
Hospitalization	20.6	21.9	0.94 (0.61, 1.31)	0.72
ED visit	36.0	39.0	0.92 (0.65, 1.31)	0.64
All cause
Hospitalization or ED visit	123.8	170.5	0.73 (0.58, 0.90)	< 0.003
Hospitalization	56.8	79.3	0.72 (0.54, 0.94)	0.02
ED visit	67.0	91.2	0.73 (0.56, 0.96)	0.02

ED, emergency department.

Results expressed as mean events per 100 patient-years.

The use of respiratory medications was significantly more common in the DM group than in the UC group for action plan medications (steroids, antibiotics), long acting bronchodilators but not for short-acting bronchodilators or inhaled corticosteroids. Quality of life as measured by the change in St. George's Respiratory Questionnaire was improved to a significantly greater extent in the DM group as compared to the UC group (1.3 vs 6.4; p < 0.001). The rate ratios for the primary outcome were better for patients who received more telephone support 0.78 (95% CI, 0.35 to 1.74; P = 0.53) for the patients with 4–8 calls vs 0–3 calls and 0.46 (95% CI, 0.22 to 0.96; P = 0.04) for patients with 9 or more calls vs 0–3 calls.

Costs associated with the DM intervention are summarized in Tables 2a and 2b. The cost of the DM intervention using 4 respiratory therapy full-time equivalents (FTE) plus printed materials and phone lines totaled $241,620 with a mean per-patient cost of $650. The cost of using 4 registered nurses instead of respiratory therapists in the DM group increased the total cost of the DM intervention to $315,716 with a mean per-patient cost of $849.

Table 2a.  Disease management program related costs using respiratory therapists

Resource	Unit Cost	Units for the 5 participating VA centers	Total Cost
Printed educational material	$1 per booklet	400	$400
Telephone and line	$420 per year	5	$2100
Respiratory therapist salary and fringe benefits	$59,780 per FTE	4	$239,120
Total			$241,620
Per patient cost			$650

Table 2b.  Disease management program costs using registered nurses

Printed educational material	$1 per booklet	400	$400
Telephone and line	$420 per year	5	$2100
Registered nurses salary and fringe benefits	$78,304 per FTE	4	$313,216
Total			$315,716
Per patient cost			$849

FTE = full time equivalent.

Costs for hospitalizations, ED visit costs, and pulmonary prescriptions in the two treatment groups over the 12-month follow-up period are summarized in Table 3. Using national average VA DRG hospital costs, mean hospitalization cost per patient was $3493 ± $4260 in the DM group compared to $4610 ± $4599 in the UC group. Mean emergency department visit cost per patient was $221 ± 557 in the DM group compared to $362 ± 729 in the UC group. Mean pulmonary prescription costs were $127 ± 139 in the DM group compared to $112 ± 157 in the UC group. The total mean healthcare costs in the DM group was $3841 ± 4678 compared to $5084 ± 5060 in the UC group representing a $1243 per-patient cost savings for the DM program. Adding the cost of the DM intervention, the total per-patient cost in the DM group was $4491 ± 4678 representing a $593 per patient total cost savings associated with the DM program. This difference in cost between the two groups was not statistically significant.

Table 3.  Mean per patient direct health care costs (U.S. Dollars) in the two treatment groups^*

	Disease Management (n = 372)		Usual Care (n = 371)
Cost Variables	Number	Cost ($)	Number	Cost ($)	Difference per Patient ($)
Hospitalizations	202	3493 ± 4260	275	4610 ± 4599	−1117
Emergency Department visits	238	221 ± 557	316	362 ± 729	−141
Pulmonary Drug Prescriptions	372	127 ± 139	371	112 ± 157	+15
Total		3841 ± 4678		5084 ± 5060	−1243

^*Mean ± SD.

A sensitivity analysis that increased the number of patients that the respiratory therapy case managers cared for to 125 patients per FTE and 150 patients per FTE allowed for a potential reduction to 3 and 2.5 FTEs to implement the DM intervention (Table 4). The corresponding per patient costs for the DM program at these increased case loads were reduced to $484 and $403, respectively. The corresponding net per patient cost savings for the increased case loads of 125 and 150 patients per FTE for the DM patients compared to the UC patients was increased to $759 and $840, respectively, compared to $593 using 93 patients per FTE. These additional differences in cost between the treatment groups were also not statistically significant.

Table 4.  Impact of respiratory therapist case load on cost in the two treatment groups

Patient caseload per respiratory therapist (No.)	Disease Management Group (n = 372) Cost per patient ($)	Usual Care Group (n = 371) Cost per patient ($)	P-value
93
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	650	0
Total Cost	4491 ± 4678	5084 ± 5060
Mean Difference	−593 (95% CI, −1856 to 678)		p = 0.52
125
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	484	0
Total Cost	4325 ± 4678	5084 ± 5060
Mean Difference	−759 (95% CI, −2022 to 512)		p = 0.44
150
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	403	0
Total Cost	4244 ± 4678	5084 ± 5060
Mean Difference	−840 (95% CI, −2189 to 409)		p = 0.39

The sensitivity analysis replacing respiratory therapists as case managers with registered nurses whose salaries are approximately 20% higher was still associated with a cost savings in the DM group compared to the UC group, but these differences were lesser than those achieved through the use of respiratory therapists (Table 5). Using nurses instead of respiratory therapists increased the cost of the DM program from $650 per patient to $849. Increasing case load for nurses to 125 and 150 patients reduced the DM intervention per-patient cost to $632 and $526, respectively. The mean per-patient cost savings using registered nurses to administer the DM program using the original caseload of 93 patients was $394. With increasing case loads of 125 patients per FTE and 150 patients per FTE, the cost savings using nurses increased to a mean per-patient savings of $611 and $717, respectively. None of these differences in total cost between the groups were statistically significant.

Table 5.  Sensitivity analysis of a nurse case management program on cost in the two treatment groups

Patient caseload per respiratory therapist (No.)	Disease Management Group (n = 372) Cost per patient ($)	Usual Care Group (n = 371) Cost per patient ($)	P-value
93
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	849	0
Total Cost	4690 ± 4678	5084 ± 5060
Mean Difference	−394 (95% CI – 1668 to 889)		0.61
125
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	632	0
Total Cost	4473 ± 4678	5084 ± 5060
Mean Difference	−611 (95% CI – 1905 to 695)		0.52
150
Healthcare Cost	3841 ± 4678	5084 ± 5060
Disease Management Cost	526	0
Total Cost	4367 ± 4678	5084 ± 5060
Mean Difference	−717 (95%CI – 2071 to 625)		0.46

DISCUSSION

In this large multicenter trial, a relatively simple, multicomponent, DM intervention program substantially reduced the mean healthcare costs in patients with severe COPD. Based on the resource allocation used in this trial which included 743 patients and hospital costs based on the VA HERC hospital costs, the per patient cost savings associated with the DM intervention depending on respiratory therapist caseload ranged from $593 to $840 compared to UC.

Previous trials of disease management in COPD have been highly heterogeneous in terms of size, type of intervention, duration of follow-up, and outcomes. Systematic reviews have reached differing conclusions about the overall value of disease management for advanced COPD, but all recognized the potential value of this type of intervention and the need for larger and more detailed studies ([16–18]). Several previous smaller trials have shown sizeable reductions in selected aspects of health care utilization, including hospitalization (6–9). The interventions in all of those studies included various combinations of multiple home visits or clinic visits, utilizing physicians or nurses or both.

Our results are consistent with the positive results obtained in those studies, even though there were important differences in our intervention program. Our intervention required only a single on-site educational session supplemented by monthly reinforcing telephone calls, all of which were performed by respiratory therapists. Although it is difficult to identify with certainty the relative contribution of the individual components of our DM program, we believe that education of patients to recognize the early symtoms of COPD exacerbation and self initiation of antibiotics and steroids as per the action plan along with phone calls by the case manager likely provided the greater benefit. Thus, we have shown that a relatively simple intervention program can be as effective as other programs outlined in the recent Cochrane Review (18). Additionally, our DM program was a practical trial that has the potential for application in a large group of high-risk COPD patients.

Of the 4 largest DM intervention trials in patients with COPD that have demonstrated reductions in healthcare resources, only 2 performed economic analyses of their interventions. In the trial by Farrero et al., 94 patients were randomized to home care or usual care. Over a one year follow-up, hospitalizations and emergency department visits were significantly reduced ([9]). Hospital costs were $48,000 in the home care group compared to $121,000 in the usual care group. Emergency room costs were $4200 compared to $13,600 in the usual care group. The cost of home care was $38,000. The total cost of home care was $90,000 compared to $136,000 in the usual care group producing a cost savings of approximately $47,000. This trial was limited by the relatively small sample size.

Bourbeau et al. randomized 190 patients to a DM intervention or usual care with a one year follow-up (6, 14). In this trial, nurse case managers handled an average of only 14 COPD patients. Nurses made visits to patients’ homes on a weekly basis for the first 6 to 7 weeks and documented real time spent with each patient. As a result, the cost of the DM program averaged $3778 per patient. The mean per patient healthcare costs over the year of follow-up were $3336 for the DM group compared to $6674 for the usual care group which represented an approximate $3300 per patient reduction in healthcare expenditures.

As a result of the high cost of their DM intervention, patients receiving DM actually incurred a $440 higher overall cost. It was only when the caseload in this study was more than tripled did the DM program result in a significant cost savings. Given the intensity of the DM intervention employed in this study, it is uncertain whether this level of workload could be easily achieved. We took a much more conservative approach in our sensitivity analysis in which the base caseload was increased from 93 patients per case manager to a high of 150 patients per case manager representing a 67% increase in workload. We believe this to be a more reasonable approach than a more than 3-fold increase in workload.

In addition, the number of hospitalizations occurring in this study was substantially higher than in our study. In the Bourbeau study, 95 patients incurred 167 hospitalizations (1.8 hospitalizations per patient) in the UC group and 96 patients incurred 91 hospitalizations (0.95 hospitalizations per patient) in the DM group. In our study, 202 hospitalizations occurred in 372 DM patients (0.5 hospitalizations per patient) and 275 hospitalizations occurred in 371 UC patients (0.74 hospitalizations per patient).

These differences would suggest that the two patient groups were different and the severity of illness was higher in the Bourbeau group. Given these differences, it is possible that our DM intervention model may not be applicable to all patients. However, given the limited data on cost savings with disease management in COPD, the results of our study demonstrate sizeable cost savings with a simplified disease management intervention that can be more generalizable and applicable in the real world setting.

Our post-hoc economic analysis had a number of limitations. A cost analysis was not prospectively planned. Hence, costs were not captured as healthcare resources were consumed. The most difficult aspect of any pharmacoeconomic analysis is the ability to accurately determine the true costs of health-care resources. We used data from the VA HERC data set and the VA DSS national database to determine costs for hospitalizations, ED visits, and pulmonary drug prescriptions. These costs are directly determined from VA databases and are probably the most accurate reflections of the direct VA costs we were attempting to identify.

Additionaly, this was a randomized non-blinded trial. Only patients in the DM intervention group were actively case managed and DM group patients called their case managers while the control group had access to the general nurse on call for the hospital. Outpatient clinic visits were not a part of the original study. Hence, clinic visits were not tracked and we could not include the costs of those visits into our cost analysis. This is probably not a major drawback, given the fact that as with many other studies of COPD DM programs, hospital costs account for the greatest proportion of overall healthcare costs (16–18).

In summary, this is the first large scale study where a relatively simple COPD DM program implemented within a regional Veterans Affairs medical system and administered by respiratory therapists reduced total healthcare costs by an average of $593 per patient over the course of a one-year follow-up. Sensitivity analyses demonstrated that increasing the case load of the respiratory therapist case managers resulted in greater cost savings. Using registered nurses instead of respiratory therapists as case managers continued to be associated with cost savings, albeit with a smaller cost savings.

Despite the cost savings, the differences in total healthcare costs between the two treatment groups were not statistically significant. This failure to demonstrate a significant difference in cost resulted in large part from the substantial variability in the costs of hospitalizations observed in the two treatment groups. Despite the lack of statistical significance in total healthcare costs, this DM intervention produced a significant reduction in hospitalizations or ED visits, an improved quality of life, and a cost savings rather than an increase in healthcare expenditures. We conclude that this COPD DM intervention is worthwhile and can be implemented in healthcare systems with large populations of high-risk patients with moderate to severe COPD and be considered with caution in patients with very severe COPD.

ACKNOWLEDGMENTS

Robert A. Petzel MD, VISN 23 Director- Current Under Secretary for Health, Department of Veterans Affairs; Janet Murphy VISN 23 Primary Care Service Line CEO-Current VISN 23 Director; Todd Wagner and Mark Smith, VA Health Economics Resource Center (HERC).

Declaration of interest

The authors report no conflicts of interest. Kathy L. Rice received funding in a grant from Wyeth. The authors alone are responsible for the content and writing of the paper.