Impact of New Antihypertensives on Healthcare Utilization by Hypertensive Patients
The authors advocate pharmacoeconomic evaluation of new antihypertensive drugs before acceptance by formulary decision makers.
The rising cost of healthcare has always been a prime concern for policymakers, healthcare economists, healthcare payers, and society at large. It has been suggested that a significant reason for this escalating cost is use of the newer and more costly medications that are introduced in the market each year. Advocates who support acceptance and use of newer drugs argue that newer medications have better clinical outcomes, tend to reduce the overall healthcare cost, and contribute to increased quality of life and longevity.1 The objective of this study was to verify the hypothesis that adoption of newer medications is associated with decreased healthcare cost and utilization of healthcare services among patients with hypertension. Our intention was to understand the relationship between new drug adoption and healthcare utilization from the patient’s perspective.
BACKGROUND
Prescription drug expenditures accounted for 13% of total US healthcare expenditures in 2006. This cost soared from $50 billion spent in 1990 to $219 billion in 2006.2 Despite strong evidence from the famous Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial supporting the superiority of thiazide-type diuretics over other groups of antihypertensive medications, both in terms of clinical outcomes and treatment costs, use of newer medications has increased more than use of diuretics.3-5 Joint National Committee guidelines recommend diuretics as the first choice in hypertensive patients with no comorbidities.6 However, differences still exist between these guidelines and the initial drug of choice in clinical practice.7 It is estimated that nearly $60 billion is spent on managing hypertension and its complications annually in the United States, yet only 34% of the patients treated with antihypertensive medications reach desired blood pressure goals.6,8 Recent studies have stated that much of this increase in cost is attributable to the introduction of new and generally more costly brand name drugs that replace older, less expensive medications.9 It is crucial to understand this current healthcare spending, as current drug expenditures have been identified as strong predictors of future nondrug expenditures.10 Policymakers, both private and public, have devised numerous ways to reject the replacement of older, less expensive drugs with newer costly drugs in the absence of proven improvement in outcomes.2 The question every healthcare provider should address before prescribing a new and expensive treatment is whether the drug is worth the cost.
Lichtenberg found that use of newer drugs reduced nondrug costs 3.9 to 7.2 times as much as it increased drug expenditures.9 This study compared the drug age with healthcare expenditures. The drug age was defined as the number of years prior to 1996 that the active ingredient was first approved by the US Food and Drug Administration (FDA). Lichtenberg used the number of prescriptions as the unit of observation and found that, on average, replacing a 15-year-old drug with a 5.5-year-old drug would save $72 in nondrug healthcare spending and cost $18 more in drug spending, resulting in a drug offset effect of $54.
In another study, Lichtenberg used the medical condition instead of prescriptions as the unit of observation and estimated a larger drug offset of $111.11 One of the major limitations of this study was that it did not take into consideration the severity of illness of the patient. Zhang and Soumerai used a generalized linear model with a log link function and gamma distribution with the Medical Expenditure Panel Survey (MEPS) 1996-1998 data and estimated $23 as the cost of new drugs after adjusting for illness severity, compared with $18 in the Lichtenberg study.2 In a study of the new cardiovascular drugs, Miller et al controlled for the number or mix of drugs and found the relationship between the new drug spending and drug age to be statistically insignificant.10 So these studies highlighted the fact that, on an aggregate level, the new drugs seemed to be worth the cost, although that may not be the case in specific disease states.
The objective of this study was to analyze the impact of prescribing a new antihypertensive drug on total healthcare expenditures, nondrug healthcare expenditures, and emergency department (ED) visits in US patients with hypertension. We examined whether adopting a new antihypertensive drug was associated with changes in (1) total healthcare expenditures, (2) nondrug healthcare expenditures, and (3) total number of ED visits. This study included certain new drugs approved by the FDA after 1996 that were not considered in previous studies.
METHODS
We performed a retrospective follow-up study using the 1999 MEPS database. Patients were defined as hypertensive when they had a diagnosis of hypertension (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 401-405). Hypertensive patients during the first round of the year (round 3 of 1998 and round 1 of 1999) who had been prescribed at least 1 antihypertensive drug during that period (first round) were included in the study (see
Figure
).
Analytic Overview
Antihypertensive drugs approved by the FDA during 1996, 1997, and 1998 were defined as new antihypertensive drugs in this study. New drug adopters were those hypertensive patients who had been prescribed at least 1
of the identified new antihypertensive medications during the first round in 1999. Non—new drug adopters were defined as those hypertensive patients in round 1 of 1999 who had not been prescribed any of the new antihypertensive drugs. Both populations were followed until the end of the year. The healthcare expenditure data at the end of the year were compared between the 2 groups. The number of unique medications for all conditions was used as a surrogate marker for the patient’s comorbidity index. Prescriptions with similar active ingredient(s), irrespective of dosage form and strength, were considered unique medications. Total healthcare expenditure, as defined in MEPS, was the sum of direct payments for care provided during the year, including out-of-pocket payments and payments by private insurance, Medicaid, Medicare, and other sources. Payments for nonprescription drugs and alternative care services were not included in the MEPS total expenditures. Indirect payments not related to specific medical events, such as Medicaid Disproportionate Share and Medicare Direct Medical Education subsidies, also were not included.12
Nondrug costs were determined by subtracting prescription drug costs from total healthcare costs. The number of ED visits was taken as one of the outcome measures. All the results were adjusted for sex, age, race, ethnicity, severity of disease, income level, and type of insurance. The type of insurance was categorized as private, public, or uninsured.
According to the Center for Drug Evaluation and Research data briefing, the FDA approved 102, 101, and 65 standard new drug applications (NDAs) in the years 1996, 1997, and 1998, respectively; of these standard NDAs, 35, 30, and 14 approvals were for standard new molecular entities for the years 1996, 1997, and 1998, respectively.13 In the years 1996, 1997, and 1998, the FDA approved 3, 7, and 4 NDAs, respectively, for treatment of hypertension.14 New drugs for this study included all new molecular entities, new molecules in previously existing drug classes, new product extensions (eg, extended/sustained-release dosage forms), and new combination drugs with an existing patent. Generic drugs and drugs with expired patents were not classified as new drugs. Hypertensive patients taking any 1 of these 14 new drugs were identified as new drug adopters. Of these 14 drugs, 7 were new molecules from previously existing drug classes, 3 were new product extensions, and 4 were new drug combinations.
Data Analysis
New drug adopters and non—new drug adopters were compared for differences in demographic characteristics, income level, and type of insurance with the χ2 test. A multiple regression analysis using a generalized linear model with a log link function and gamma distribution was used to analyze the association between new drug adopters and healthcare expenditures, adjusting for age, sex, race, ethnicity, income, severity of disease, and insurance type. Interpretation was done by taking the natural exponential [exp(β1)] of the regression coefficient for the expenditure variables. The association between new drug adoption and ED visits was analyzed using the ordinary least squares regression method. The variables that were compared and reported were total healthcare cost, nondrug cost, and total number of ED visits. All analysis was performed without any reference to the sample weights. The study sample was treated as simple random rather than stratified random. The level of statistical significance was P ≤.05, and all statistical computations were done with SAS 9.1 software (SAS Institute Inc, Cary, NC).
RESULTS
Of 24,618 unweighted observations, 2721 (11.05%) represented patients who had been diagnosed with hypertension (ICD-9-CM codes 401-405) in the 1999 MEPS database. A total of 1306 (47.99%) patients had been diagnosed as hypertensive in the first round of the 1999 MEPS survey; 1149 (87.97%) had been prescribed at least 1 antihypertensive medication and formed the final population for this study. Sixty-three patients (5.48%) were prescribed new antihypertensive drugs and hence grouped as new drug adopters (
Table 1
).
No significant differences were observed between new drug users and non—new drug users in terms of age, race, ethnicity, income, and insurance status, although new drug users were significantly more likely to be female (
Table 2
). The majority of new drug users and non—new drug users (60.32% and 64.27%, respectively) were younger than age 45 years. More than a third in both groups had an annual income below $20,000. About 79% of the new drug adopters were insured versus 88% of the non–new drug users. Among insured patients, 80.0% of new drug adopters and 80.33% of non–new drug adopters had private insurance. Both groups were dominated by Caucasians, mostly white non-Hispanics; 77.78% and 78.55% of the new drug adopters and non–new drug adopters, respectively, were whites.
The results of the regression analysis showed that the total healthcare expenditure and the nondrug expenditure were significantly associated with adoption of new antihypertensive drugs (
Table 3
). Upon converting these values into their natural exponential form, the total healthcare expenditures and the nondrug healthcare expenditures of new drug adopters were found to be 1.63 (P = .019) and 1.79 (P = .013) times higher, respectively, than the corresponding expenditures of the non—new drug adopters (
Table 4
). Though the new drug adopters had more ED visits than non—new drug adopters, this difference was insignificant (P = .089). Age and income were significantly associated with total and nondrug expenditures. Hypertensive patients older than age 64 years had a 2-fold increase (P ≤.001) in the total healthcare expenditure and about a 1.7-fold (P ≤.001) increase in the nondrug expenditure compared with patients age 44 years or younger. Hypertensive patients age 45 to 64 years had total and nondrug healthcare expenditures 1.6 times (P ≤.001) and 1.39 times (P = .011), respectively, greater than those of patients age 44 years or younger. The total and nondrug expenses of patients with annual incomes between $50,000 and $100,000 were 41% (P = .002) and 42% (P = .004) lower than those of patients with an annual income lower than $20,000. Those with annual incomes higher than $100,000 spent almost 52% (P = .017) and 60% (P = .007) less in terms of the total and nondrug expenses, respectively, than patients with an annual income lower than $20,000.
African American patients had lower total and nondrug healthcare expenses than the Caucasians (0.16 times lower [P = .01] and 0.18 times lower [P = .027], respectively). Sex and ethnicity had no significant association with total and nondrug healthcare expenditures. Patients with public insurance spent 4.81 times (P ≤.001) and 4.44 times (P ≤.001) more, whereas those with private insurance spent 2.10 times (P = .002) and 1.94 times (P = .027) more than those patients without any healthcare insurance in terms of total and nondrug costs.
DISCUSSION
In this study, the association between adopting a new drug and increased healthcare expenditure was found to be significant. Results showed that patients taking new antihypertensive drugs had higher annual total healthcare costs and nondrug costs, and more ED visits than patients taking old antihypertensive drugs; however, only the total cost and nondrug cost were statistically significant. So this study refutes the argument that new medications reduce nondrug costs in the short run, as well as contradicting the argument that in real-world settings new antihypertensive drugs reduce ED visits compared with old medications, as concluded in some studies.2,9 Because new drugs definitely cost more than the older medications, an increase in the total cost is understandable. However, the increase in the nondrug cost among the new drug adopters counters the argument that new drugs always are cost-effective.
The increase in the nondrug cost could be explained by the increase in healthcare utilization. The results show new drug adopters have more ED visits, which though statistically insignificant could have contributed to the higher nondrug cost. Similarly, new drug adopters also may have utilized more healthcare services. In this study all drugs newly approved by the FDA in 1996-1998 were defined as new drugs, including all new drug molecules, new product extensions, and new drug combinations still in their patent period. It is possible that these new drugs still had a high price because of ongoing or extended patent protection, but clinically they may have limited or marginal benefits compared with the older, off-patent drugs. Of the 14 new drugs considered in this study, 7 were product extensions or new combinations rather than new chemical entities, which could have influenced the study findings. We know that product extensions and new combinations are some of the successful strategies used by companies to counter generic entry in the postpatent period.15 Of the 1035 new drugs approved by the FDA between 1989 and 2000, 674 (65%) were modified versions of the existing drugs, according to the National Institute for Health Care Management.16 The strategy of introducing product line extensions and new combinations has been quite successfully used to shift consumers from expired/soon-to-expire original formulations.17 We acknowledge that new breakthrough molecules are cost-effective, but that may not be true for drugs that are mere structural modifications of existing molecules.
Nonadherence and compromised behavior are a few of the factors that deter drug treatment in real-world scenarios.18 It is possible that the new drug adopters in this study had suboptimal drug adherence rates. The other factor that might explain these study findings is the lifestyle behavior of patients adopting new antihypertensive medications. Patients on new medications may have or been given an impression that the new drug will take care of their high blood pressure, and this perception enticed them to compromise the behavior modifications required as an adjunct to their drug treatment to achieve the desired outcomes. This study was adjusted for health condition (comorbidity); hence, the notion that new drug adopters had worse health was not supported. But the possibility that the new drug adopters had more severe hypertension than the non—new drug adopters cannot be denied.
Study patients with government-sponsored health insurance spent more than the patients with private insurance, possibly signifying overutilization of healthcare services. This difference also may be attributable to different reimbursement rates for similar services and/or products, or utilization of more costly services. Also, publicly insured hypertensive patients had more ED visits than privately insured and uninsured populations. People with higher incomes (>$50,000 per year) had the lowest total healthcare and nondrug expenditures and even fewer ED visits. Perhaps because they were affluent, they had access to and the ability to avail themselves of healthcare services and hence enjoyed better health than low-income patients. The study also found that patients older than age 64 years had higher healthcare costs, a finding supported by previous studies.19
This study had certain limitations that may have influenced the results. MEPS data lacked information on the severity of disease. Although the results were adjusted for the comorbidity index, availability of clinical measures such as the blood pressure of the patient would have been helpful. MEPS data provide cross-sectional information; hence, the analysis was done with the assumption that all study participants had similar healthcare utilization prior to the study period. Also, new drugs were defined as those drugs approved by the FDA during the years 1996, 1997, and 1998, and the study used the 1999 MEPS data. So the new drugs approved in 1996 had 3 years of effective time in the market, whereas those approved in 1997 and 1998 had 2 years and 1 year, respectively, as the effective time in the market to analyze visible impact. This difference in the effective market presence may limit the exactness of study findings.
CONCLUSION
Overall, adopting new antihypertensive drugs was associated with higher healthcare cost and healthcare utilization among hypertensive patients. New drugs are cost-effective when they prevent the use of more expensive healthcare resources. Certainly they are costeffective when they come as breakthroughs. However, the cost-effectiveness of new drugs has been challenged when new drugs are line extensions of existing drugs that had been successful in the market. In fact, this study sample mostly consisted of extensions rather than new breakthroughs.
Further research is required to understand the exact relationship between the adoption of new antihypertensive medications and healthcare cost. The results of this study essentially suggest that prescribers of new drugs should exhibit caution before acceptance. The results also signify the need to scrutinize the pharmacoeconomic evaluation of each new drug before acceptance by formulary decision makers, payers, physicians, and pharmacy managers.
