Ozminkowski RJ, Wang S, Marder WD, Azzolini J, Schutt D. Cost implications for the use of inhaled anti-inflammatory medications in the treatment of asthma. Pharmacoeconomics 2000;18:253-264.
Overall Study Question
The objective of this study was to compare the expected costs of treating patients with asthma with, versus without, inhaled anti-inflammatory medications (i.e. corticosteroids, sodium cromoglycate, nedocromil), while attempting to adjust for other factors that also influence medical care expenditures. The study population was comprised of 7,466 patients continuously enrolled in one of 35 health plans in the United States in 1995 and 1996. These health plans were offered by 10 large employers who contributed inpatient, outpatient, pharmaceutical and health plan enrollment data to the MEDSTAT Group Marketscan Family of databases. Therefore, this study is based entirely on administrative health plan data. The effect of inhaled anti-inflammatory medications on two outcomes was evaluated: total inpatient, outpatient, and pharmaceutical expenditures, and asthma-related expenditures, in the 1996 calendar year. Simulation models were used to determine the differences in costs assuming all patients were, or were not, treated with an anti-inflammatory medication.
Are the Results of the Study Valid?
1. Did the analysis provide a full economic comparison of health care strategies?
This study does not provide a full comparison of economic strategies, and does not calculate any incremental costs associated with any particular treatment. Outcomes were not measured. The authors suggest that this study was done “from the perspective of the payer, and to some extent, the patient”. They suggest that the perspective of the patient is taken, given that they include any co-payment and deductible in their determination of costs. In essence, this is a budget impact study, however, the inclusion of of any co-pay and deductible amounts in the “costs” deems it an inaccurate assessment of the true budget impact from the perspective of the payer. Furthermore, from the perspective of the patient, this approach captures only a fraction of the costs borne by the patient.
2. Were the costs and outcomes properly measured and valued?
Outcomes were not valued, nor differences in outcomes between treatment groups (i.e. hospital admission, emergency visits, etc.) valued or assessed. The authors do not explicitly state how “costs” were determined. However, it can be assumed that “costs” were derived from the MarketScan databases, and therefore, insurance claims were used to determine all “costs” associated with treatment. However, these claims databases are unlikely to measure costs, but rather more likely track charges. Charges may differ significantly for similar patients, depending on the patient’s insurance provider. Therefore, it would appear that this study was actually measuring charges as opposed to costs, but it is not clear whether or not these charges were consistent across all 35 health plans.
3. Was appropriate allowance made for uncertainties in the analysis?
Data were analyzed using nonlinear exponential regression to estimate the relationships between medical care expenditures in 1996 and the use of inhaled anti-inflammatory drugs. Mean and median costs, as well as standard deviations are reported. Cost differences between patients treated and not treated with inhaled anti-inflammatory medications were tested for statistical significance, as was the model simulating all patients treated versus not treated with an inhaled anti-inflammatory medications. Although standard deviations on all costs were reported, sensitivity analysis was not performed.
4. Are estimates of costs and outcomes related to the baseline risk in the treatment population?
Sub-group analyses were not performed. Although Adjusted Clinical Group (ACG) was used to control for disease severity in the multivariate regression analysis, costs associated with treating patients with different disease severities was not reported. The effects of drug therapy on the different costs (i.e. inpatient, outpatient, or pharmaceutical costs) were not provided.
What were the Results?
1. What were the incremental costs and outcomes of each strategy?
Incremental costs were not calculated. All costs were valued in US dollars. The mean total health care expenditure in patients treated with inhaled anti-inflammatory drugs was $4,534.08 (median $2,300.91; SD $8,827.56) versus $3,897.54 (median $1,329.04; SD $13,187.99) for patients not receiving an anti-inflammatory medication. The difference in mean cost was not statistically significant, however the median cost was significantly lower in those patients who did not receive an inhaled anti-inflammatory medication. Analysis of asthma-related costs yielded similar results. The median asthma-related cost was $453.23 for anti-inflammatory treated patients versus $99.63 for patients not treated with an inhaled anti-inflammatory medication (p=0.01 ). However, those patients not treated with anti-inflammatory medications (n=2,242) differed significantly from those treated with anti-inflammatory medications on a number of variables, including type of insurance plan, age, geographic region, and the primary policy holder in the family. This would suggest that these two groups are systematically different, and therefore any differences in costs of disease management cannot be attributed solely to differences in use of anti-inflammatory medications.
The results of the regression modeling and simulation showed that the estimated mean total health care expenditure per person, if all patients in the sample received an inhaled anti-inflammatory, would be $2,954.96, versus $3,899.78 if no patients were treated with inhaled anti-inflammatory drugs. Thus, overall health care costs would decrease by $944.82 per patient. Conversely, when only asthma-related expenditures were evaluated, if all patients were treated with an inhaled anti-inflammatory, total asthma related costs were estimated to be $940.36 per patient versus $441.74 per patient if nobody was treated with an inhaled anti-inflammatory. Therefore, asthma management with inhaled anti-inflammatory drugs would result in $498.74 greater expenditure per patient.
However, as discussed in the first analysis, those treated and not treated with anti-inflammatory drugs are not comparable on a number of different variables. Although these variables were included in the multivariate regression, it is likely that patients treated versus not treated with inhaled anti-inflammatory drugs were systematically different in other drug management and cost of care aspects that were not measured and may have been, in fact, immeasurable. Further support of this is provided by the assessment of model fit, particularly for the regression determining relationships with asthma-related health costs. The final model resulted in an adjusted r-square of only 0.038. Although the model was significant, this suggests that this model describes very little of the variance associated with asthma-related health care costs.
2. Do incremental costs and outcomes differ between sub-groups?
Sub-group analyses were not performed.
3. How much does allowance for uncertainty change the results?
Although results of any sensitivity analyses were not provided, the variance of the total and asthma-related costs was large. The standard deviations for the mean total and the asthma-related costs for patients treated with anti-inflammatory medications were $8,827.56 and $2,275.25, respectively. The corresponding standard deviations in costs for patients not treated with anti-inflammatory medications were $13,187.99 and $1,930.55. This suggests two plausible outcomes: that there was the potential for the results to be reversed, and that there was significant heterogeneity within the groups in terms of total and asthma-related health costs. This provides further evidence of systematic differences between the groups.
Will the results help me in caring for my patients?
1. Are the treatment benefits worth the harms and costs?
The results suggest dominance of using anti-inflammatory medications, at least in terms of total health care costs. Outcomes were not measured, so harm could not be assessed. The perspective of this analysis and the limitations of the data as outlined previously limit the applicability of the results. The true costs of the administration of inhaled anti-inflammatory drugs thus remains to be determined.
2. Could my patients expect similar health outcomes?
Outcomes were not measured.
3. Could I expect similar costs?
These data were derived from administrative health care records, and therefore more likely reflect charges as opposed to costs. Furthermore, this was a US-based study of 35 health insurers; charges for similar procedures or medications may have differed between insurers, but this was not delineated in the publication. The inclusion criteria for the study limited the eligibility to patients to those who were likely to have more severe disease. Therefore, these results cannot be extrapolated to any other setting. Considering all study limitations, the true impact of using inhaled anti-inflammatory medications on both total costs of health care, and asthma-related costs, remains in question.
It is well established that inhaled anti-inflammatory medications, particularly corticosteroids, should be the mainstay of therapy in all but the mildest asthmatics. This is based on ample evidence of improved outcomes in patients treated with appropriate doses of inhaled corticosteroids compared to those who continue to rely on short-acting beta-agonists. Clinical improvements associated with appropriate use of inhaled corticosteroids include reduction in symptom severity, improvement in peak flow and other objective spirometric measures of lung function, decreased airway hyperresponsiveness, prevention of exacerbations, and possibly the prevention of airway remodeling. These benefits are expected to be accompanied by increased drugs costs; other outpatient and inpatient costs are expected to decrease while other economic implications, including quality of life and indirect costs such as time off work, are expected to improve. This study did not stratify costs to allow for the determination of whether increased drug costs resulted in decreased inpatient and outpatient costs.
The primary methodologic limitations of this study include:
1) the heterogeneity of patients receiving and not receiving inhaled anti-inflammatory medications;
2) the use of charges versus costs, and the lack of providing explicitly how costs were determined;
3) inclusion criteria limited to patients who likely have moderate to severe disease;
4) a poor model fit in the determination of factors associated with asthma related costs; and
5) the lack of breakdown of inpatient, outpatient, and pharmaceutical costs associated with each treatment.
Therefore, the apparent greater asthma-related costs associated with inhaled anti-inflammatory use remains questionable.
Given the known benefits of inhaled corticosteroid administration and their inclusion as the mainstay of therapy in the current asthma management guidelines, this study should in no way impact the decision of whether or not to prescribe an inhaled anti-inflammatory medication.