Abstract
Purpose
Practice patterns were investigated for an insured population to determine if prescribing patterns, switching, and relative adherence by drug class for first-line antihypertensive medications adhered to national guidelines.
Procedures
Drug use was obtained from pharmaceutical claims. Prescriptions were categorized into 6 drug classes for analyses. Adherence with antihypertensive medications was based on a medication possession ratio or 0.8 or greater. For the analyses, 28,073 patients were categorized into groups: hypertension alone, hypertension plus diabetes, and hypertension plus congestive heart failure. Patient and physician characteristics affecting prescribing, switching, and adherence were analyzed using multivariable logistic regression analysis.
Findings
Thiazide diuretics were used and adhered to less often, despite national guideline recommendations. New drug classes were used more highly.
Conclusions
Inconsistency exists between guidelines and practice as older, cheaper drugs were used less and more expensive drugs were used more often with better adherence.
Introduction
Hypertension as a leading medical risk factor for heart disease and stroke1 affects approximately 65 million Americans.2–4 Appropriate treatment and medication use might substantially reduce medical risks of hypertension. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC 7) recommends for most persons with uncomplicated hypertension an initial use of a thiazide-type diuretic alone or in combination with another agent such as an ACE Inhibitor, a beta blocker, an angiotensin receptor blocker (ARB), or calcium channel blocker.1
Patients with some disease states and certain ethnicities may have compelling indications requiring additional antihypertensive agents. For instance, the JNC 7 guidelines recommend persons with heart failure use a thiazide-like diuretic, a beta blocker, an ACE Inhibitor, an angiotensin receptor blocker (ARB), or an aldosterone antagonist; but that these persons not use a calcium channel blocker. For persons with diabetes, the guidelines recommend a thiazide-type diuretic, a beta blocker, an ACE Inhibitor, an ARB, or a calcium channel blocker, but not an aldosterone antagonist.1
In the Holmes et al4 study, the authors reported that prescribing practices of health providers did not match guideline recommendations. Nearly half of the individuals with hypertension and various co-morbidities did not receive first-line treatments. The authors further described substantial social and financial costs of prescribing trends using newer drugs rather than less expensive traditional drugs, like diuretics and beta blockers, and without evidence that the newer drugs offered superior benefits. Holmes et al noted a lack of comparison studies with detailed, individual level prescription information on antihypertensive agents. Recent evidence suggest, on the other hand, that adherence with some newer agents may be better, suggesting a preventive advantage.5
In Hawai'i 1 out of 3 deaths is a result of hypertension-related problems.6Native Hawaiians have exceptionally high risks. This study examines antihypertensive prescribing patterns for the initial drug prescribed and adherence among an ethnically diverse group of patients enrolled in a large health plan in Hawai'i.
Methods
The design of this study was a non-concurrent longitudinal study analyzing administrative data provided by a large insurer in the state of Hawai'i. A limited data set, included dates of the first prescription, was created for the study. Individual participants could not be identified from this data. An exemption from IRB review was granted prior to initiating the study by the University of Hawai'i Institutional Review Board.
This study population consisted of members with a diagnosis of hypertension based on ICD-9 codes and who were prescribed their first prescription for an antihypertensive medication between the years 2002 and 2004. A minimum of 1 year of continuous enrollment before the first prescription was required to participate. Ninety-two percent of the participants had been continuously enrolled for 2 years or more. All insured members 18 years and older were eligible if they met the diagnosis and enrollment criteria.
Antihypertensive medications were identified by specific therapeutic class codes. Adherence to a prescribed medication was based on a “medication possession ratio” (the percentage of days with a prescribed medication).7 A ratio of 0.8 or greater was considered adherent for this study. The maximum allowable gaps between prescription fills were based upon possession ratios, calculated as follows: Possession ratio = days supplied for first prescription (Rx) (fill date of second Rx - fill date of first Rx). A claim separated from a previous claim with a possession ratio of 0.8 or greater was considered adherent. Adherence was examined for 6 drug classes among members who were classified into 3 disease categories: hypertension alone, hypertension with diabetes mellitus, and hypertension with congestive heart failure. Diabetes and congestive heart failure were defined by the insurer's criteria for its disease management program.
Explanatory variables for the study population included: demographic characteristics, island of residence, morbidity, insurance type, and the specialty of prescribing physician. The morbidity index used comes from the Johns' Hopkins'Adjusted Clinical Group case-mix adjustment system, which categorizes patients' clinical conditions from ICD-9 diagnoses into 1 of 6 integer categories ranging between 0 and 5. Higher numbers indicate worse morbidity. This measure of morbidity is a risk adjustment tool that measures the illness burden of patients and their expected consumption of health services.8
In descriptive analysis, frequencies were calculated for demographic variables, drug classes used, and physician specialties. Drug adherence during the first year of use was analyzed using multivariable logistic regression analysis. In analyses modeling physician specialties, members were identified as clustered within the practice of their initial prescribing physicians. These analyses were performed using generalized linear mixed effects models9 and were adjusted for patient characteristics (age, gender, isle of residence, insurance type, morbidity, and if the patient had a diagnosis of diabetes or congestive heart failure). Odds ratios compare specialists to primary care physicians on the likelihood of prescribing the listed drug class relative to thiazide diuretics.
Results
Calculations from 2002 to 2004 showing that of 28,073 patients receiving first prescriptions for an antihypertensive medication, 63% were over 50 years of age while 23% were 65 years of age or older (Table 1). The proportion of women to men was about even. Seventy percent of the participants lived on the island of O'ahu in Hawai'i; the remainder lived on rural, outer Hawaiian Islands. A substantial majority (84%) were enrolled in fee for service plans as opposed to health maintenance organizations. About 30% had a high morbidity based upon an assessment of their insurance claims in the past year. Fourteen percent of the patients with an initial prescription for an antihypertensive medication were diagnosed with diabetes and about 2% were diagnosed with congestive heart failure.
Table 1.
Demographic characteristics of study population Total: 28,073
| CHARACTERISTIC | DESCRIPTION | FREQUENCY (PERCENT) |
|---|---|---|
| Age | 20–34 | 1,701 (6.1%) |
| 35–49 | 8,801 (31.3%) | |
| 50–64 | 11,023 (39.3%) | |
| 65–79 | 5,099 (18.2%) | |
| 80+ | 1,449 (5.2%) | |
| Gender | Female | 13,652 (48.6%) |
| Male | 14,421 (51.4%) | |
| Island of Residence | O'ahu | 19,555 (69.7%) |
| Hawai'i | 3,803 (13.5%) | |
| Maui | 2,062 (7.3%) | |
| Other | 2,653 (9.4%) | |
| Insurance Type | Health Maintenance Organization | 4,492 (16.0%) |
| Fee-for-service | 23,581 (84.0%) | |
| Morbidity | Low | 19,691 (70.1%) |
| High | 8,382 (29.9%) | |
| Disease Category | Hypertension | 23,488 (83.7%) |
| Diabetes | 3,972 (14.1%) | |
| Congestive heart failure | 613 (2.2%) |
Of the participants with hypertension alone, 65% were prescribed ARBs or ACE Inhibitors as their initial antihypertensive medication (Table 2). Twenty percent were prescribed beta blockers. Half of the patients with diabetes or congestive heart failure received ACE Inhibitors. An additional 31% of those with diabetes received ARBs. Of the patients with congestive heart failure, about 18% were prescribed ARBs; a similar percentage was prescribed beta blockers. Calcium channel blockers were prescribed almost twice as often as thiazides to patients with congestive heart failure.
Table 2.
Percentage of first prescriptions for afltypgrtdre to metfcafeisfe drug class and disease category
| DRUG CUSS | Hypertension | Diabetes Mellitus | Congestive Heart Failure |
|---|---|---|---|
| Angiotensin receptor blockers (ARB) | 29.3% | 30.8% | 17.5% |
| ACE Inhibitors (ACE 1) | 25.6% | 50.5% | 49.9% |
| Beta blockers (BB) | 19.8% | 7.4% | 17.8% |
| Calcium channel blockers (CCB) | 11.8% | 5.2% | 9.0% |
| Thiazide diuretics (THZ) | 10.9% | 4.0% | 5.1% |
| ACE Inhibitors and CCBs | 2.6% | 2.1% | 0.8% |
About 60% of the prescriptions from primary care physicians were for ARBs or ACE Inhibitors. Beta blockers were prescribed half as often. Cardiologists and endocrinologists exhibited somewhat contrasting prescription practices. Cardiologists most frequently prescribed beta blockers followed by ACE Inhibitors and then ARBs. Endocrinologists were the reverse: they most frequently prescribed ARBs followed by ACE Inhibitors and then beta blockers. Physician prescribing patterns were further compared in regression models adjusting for patient characteristics. After adjustment, cardiologists and endocrinologists were 3 times as likely as primary care physicians to prescribe ARBs than as to prescribe thiazides. Cardiologists were also significantly more likely to prescribe ACE I inhibitors, beta blockers, and calcium channel blockers. Of all physician specialties, those who prescribed antihypertensive medications to the most patients were the least likely to prescribe thiazide diuretics (Table 3). All comparisons were statistically significant (p<0.05), although the relative odds of prescribing were moderate in magnitude. The median number of patients prescribed was 7, but a third of the physicians prescribed to 20 or more patients, and 20% prescribed to 40 or more patients. The strongest association with the number of patients treated was for ARBs: for every additional 20 patients treated, the odds of prescribing ARBs relative to thiazide diuretics increased 20%.
Table 3.
Odds ratios for prescribing listed drug classes relative to thiazide.dluretics by the number of patients to which the physician had prescribed antihypertensive medications.a
| DRUG CLASS | ODDS RATIO PER 20 PATIENTS PRESCRIBED |
|---|---|
| Angiotensin receptor blockers | 1.2 (1.2, 1.3) |
| Ace Inhibitors | 1.0 (1.0, 1.1) |
| Beta blockers | 1.1 (1.0, 1.1) |
| Calcium channel blockers (CCBs) | 1.0 (1.0, 1.1) |
| Ace Inhibitors and CCBs | 1.1 (1.1, 1.2) |
Regression models treated patients as grouped within the physicians who prescribed their initial antihypertensive medication. Analyses were adjusted for patient age, gender, isle of residence, insurance type, morbidity, and if the patient was diagnosed with diabetes or congestive heart failure.
When looking at all of the drug classes members used in their first year of treatment, most patients reported using a single drug class (Table 4). ARBs were the most common drug class for patients with hypertension alone, whereas ACE inhibitors were most frequently used by patients with hypertension complicated by diabetes or congestive heart failure.
Table 4.
Drug classes or combinations of class prescribed during first year of antt-hypertenave medication use by disease category.a
| DISEASE CATEGORY | DRUG CLASSES | PERCENT |
|---|---|---|
|
| ||
| Hypertension alone | Angiotensin receptor blocker | 25.9% |
| Beta blocker | 18.5% | |
| ACE 1 inhibitor | 18.2% | |
| Calcium channel blocker | 10.4% | |
| Thiazide diuretics | 10.4% | |
| ACE 1 inhibitor & calcium channel blocker combination | 2.6% | |
| Ace 1 inhibitor & beta blocker | 2.3% | |
|
| ||
| Diabetes | ACE 1 inhibitor | 31.9% |
| Beta blocker | 16.2% | |
| Angiotensin receptor blocker | 12.9% | |
| Ace 1 inhibitor & beta blocker | 8.5% | |
| Calcium channel blocker | 7.8% | |
| Thiazide diuretics | 4.1% | |
| ACE 1 inhibitor & Angiotensin receptor blocker | 3.9% | |
|
| ||
| Congestive heart failure | ACE 1 inhibitor | 39.9% |
| Angiotensin receptor blocker | 27.3% | |
| Beta blocker | 6.8% | |
| ACE 1 inhibitor & Angiotensin receptor blocker | 5.5% | |
| Calcium channel blocker | 4.4% | |
| Thiazide diuretics | 3.8% | |
Table lists all drug classes or combinations used by 2% or more of patients in a disease category. Ampersand separates different drug classes used by patients during year, and not necessarily at the same time.
During their first year of use and depending upon the initial drug used, 3–9% of patients switched drug classes. ARBs and beta blockers were switched least often, whereas the combination of ACE Inhibitors and calcium channel blockers were most commonly switched. Comparing the 2 most established drug classes, beta blockers were switched half as often as thiazides.
Drug adherence relative to thiazide diuretics during the first year of antihypertensive medication use was highest for ARBs, ACE Inhibitors, and beta blockers (Table 5). Among patients with diabetes, adherence was greatest with ARBS; adherence was greatest with ARBs among patients with hypertension alone or hypertension and congestive heart failure. Thiazide diuretics invariably exhibited the worst adherence.
Table 5.
Percent adherence with antihypertensive medications and odds ratios comparing adherence of other drug classes to thiazide diuretics by disease categorya
| DRUG CLASS | HYPERTENSION ALONE | HYPERTENSION & DIABETES | HYPTERTENSION & CONGESTIVE HEART FAILURE | |||
|---|---|---|---|---|---|---|
| % adhered | ODDS RATIO (95% Cl) | % adhered | ODDS RATIO (95% Cl) | % adhered | ODDS RATIO (95% Cl) | |
| Angiotensin receptor blockers | 55.0% | 2.4 (2.2, 2.6) | 41.2% | 3.1 (2.1, 4.4) | 52.3% | 3.8 (1.5, 9.8) |
| Beta blockers | 46.2% | 1.7 (1.5, 1.9) | 44.9% | 1.6 (1.1, 2.4) | 39.4% | 2.3 (0.9, 5.8) |
| Ace 1 inhibitors and CCBs | 43.3% | 1.5 (1.3, 1.8) | 41.2% | 1.4 (0.8, 2.5) | 20.0% | 1.0 (0.1, 0.8) |
| Ace 1 inhibitors | 42.0% | 1.4 (1.3, 1.6) | 43.4% | 1.6 (1.1, 2.2) | 45.1% | 3.0 (1.2, 7.2) |
| Calcium channel blockers (CCBs) | 41.3% | 1.4 (1.2, 1.5) | 37.2% | 1.2 (0.8, 1.8) | 36.4% | 2.0 (0.7, 5.6) |
| Thiazide diuretics | 33.4% | 1.0 | 33.1% | 1.0 | 22.6% | 1.0 |
Odds ratios were adjusted for age, gender, isle of residence, insurance type, morbidity, and whether the patient had coronary artery disease.
Discussion
In this Hawai'i study, the newer ARBS and ACE Inhibitors were the most common antihypertensive medications prescribed to patients not known to have been previously treated. This finding was consistent for members with hypertension alone or hypertension in combination with diabetes or congestive heart failure. In apparent contradiction to hypertension guidelines, the older beta-blockers and diuretics were prescribed less frequently than the newer drug classes. These results confirm what other investigators such as Holmes et al4 have observed: that recommended, first-line antihyper-tensive agents are not being consistently prescribed to patients with hypertension.
In the current study primary care physicians and endocrinologists wrote prescriptions most often for ARBs and ACE Inhibitors. Cardiologists most often prescribed beta blockers and ACE Inhibitors. Patients adhered best with ARBs followed by beta blockers or ACE Inhibitors, depending upon the patient's co-morbid conditions. The choice of ARBS as a first-line therapy is likened to the results of Greving, Denig, van der Veen, et al10 in The Netherlands, where angiotensin II receptor blockers were found to be first-line rather than second-line therapy for hypertension.
Reports on best practices vary and results from randomized trials differ in conclusions regarding the most effective antihypertensive medications. For instance, Psaty, Smith, Siscovick, et al11 concluded that beta-blockers and diuretics were superior as first-line antihypertensive agents to ACE Inhibitors and calcium channel blockers. In a later analysis, Psaty, Lumley, Furberg, et al12 concluded that low-dose diuretics were the most effective first-line therapy in preventing cardiovascular disease and suggested clinical practice should follow suit. In 2002, the ALLHAT Collaborative Research Group found that low dose thiazide diuretics were superior as first-line therapy and were unsurpassed in lowering blood pressure and with regard to consideration of clinical events, cost and tolerability.13 In 2007, Einhorn, et al14 concluded that for prevention of heart failure in high-risk hypertensive patients, thiazide diuretics should be the preferred first-step therapy
Following national guidelines may yield substantial cost savings. According to a simulation by Fischer & Avorn15 the cost difference between antihypertensive agents actually prescribed and those suggested by evidence-based guidelines could have saved an estimated $11.6 million. That study concluded that a vast consideration of resources are being expended on the newer and more expensive drug classes; and without clear evidence from randomized trials that patients benefit from these classes over and above older, less expensive alternatives. Free pharmaceutical samples and type of insurance may influence selection of non-first line prescriptions16 as might education, costs of treatment, and updates on clinical trials.17–18
The results of this study suggest that current practice in Hawai'i does not follow the recommended guidelines since thiazide diuretics were one of the least used drug classes. Most often clinicians started patients diagnosed with hypertension on ARBs, which over the first year of use had better drug adherence. It is possible factors such as undesirable side effects from diuretics may have influenced provider selection of drugs and patient adherence; however, that cannot be confirmed by this study. Some believe that for about two-thirds of people with hypertension and especially those with co-morbidities such as diabetes mellitus, 2 drug classes may be required to provide blood pressure control. In this study of initial drug use for hypertension, the use of multiple drug classes was much less frequent.
The results reported in this article should be interpreted with a number of limitations in mind. First, prescriptions were from billing data and represent filled prescriptions and not actual medications taken. As a consequence, adherence represents refilling of prescriptions, not necessarily the continued taking of antihypertensive medications. Second, data on free samples are unavailable. As a consequence, the medication identified as the first prescription may not necessarily be the first medication taken. Also, since our data do not report blood pressure measurements or levels, we cannot examine the effectiveness of the various medications. This could be evaluated in future studies along with numbers of patients being prescribed 2 or more antihypertensive drugs since drug combinations were excluded from the data.
Although JNC 7 guidelines and pharmacological treatment options are reasonably broad and open to interpretation the observed prescribing patterns in this study appear inconsistent with the intent of the guidelines. ARBs and ACE Inhibitors were used most often for all disease categories and thiazides were used least often. In a 2005 report, the American Journal of Managed Care5 notes results similar to these: evidence of increased prescriptions of the newer classes of antihypertensive agents due to better adherence. The report attributes adherence issues with tolerability for patients coupled with reduced side effects. Sustained therapy may afford the greatest protective cardiovascular benefits.
Other factors such as patient lifestyle modifications and preferences, efficacy of individual patient plans, costs, use/abuse of resources, side effects, safety/comfort, risks/benefits, co-morbidities, gender, and drug adherence must also be factored into the equation to help patients achieve optimal outcomes. Investigating the relative risks and benefits of available antihypertensive medications and whether optimum blood pressure control is being achieved using the various drug classes is essential. Together with these factors, approaching an understanding of why providers choose various treatments for hypertension is of utmost importance if we are to reduce the risks of heart disease and stroke related to hypertension while conserving valuable resources.
Acknowledgments
This research study was supported by funding from NIH Grants Number 1R25RR019321 Clinical Research Education and Career Development (CRECD) in Minority Institutions. This project was done through the University of Hawai'i and Hawai'i Medical Service Association, an independent licensee of the Blue Cross and Blue Sheild Association.
References
- 1.Chobanian AV, Bakris GL, Black HR, et al. National Heart, Lung, and Blood Institute Joint Nation; Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure and the Nation. High Blood Pressure Education Program. The Seventh Report of the Joint National Committee o Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–72. doi: 10.1001/jama.289.19.2560. [DOI] [PubMed] [Google Scholar]
- 2.American Heart Association, Inc. AHA Statistical Update. Heart Disease And Stroke Statistics - 200 Update: A report from the American Heart Association Statistics Committee and Stroke Statistic Subcommittee. Circulation. 2006;113:e85–e151. doi: 10.1161/CIRCULATIONAHA.105.171600. DOI:10.1161/CirculationAHA.105.171600. [DOI] [PubMed] [Google Scholar]
- 3.Guo JD, Liu GG, Christensen DB, et al. How well have practices followed guidelines in prescribin antihypertensive drugs: The role of health insurance. Value in Health. 2003;5:18–28. doi: 10.1046/j.1524-4733.2003.00212.x. [DOI] [PubMed] [Google Scholar]
- 4.Holmes JS, Shevrin M, Goldman B, et al. Translating Research into practice: Are physicians followin evidence-based guidelines in the treatment of hypertension? Med Care Res Rev. 2004;61:453–473. doi: 10.1177/1077558704269501. [DOI] [PubMed] [Google Scholar]
- 5.Reports from The American Journal of Managed Care After the diagnosis: Adherence and Persistenc with Hypertension Therapy. AmJManag Care. 2005;11:S395–399. [PubMed] [Google Scholar]
- 6.Hawaii Department of Health (DOH) Health & Vital Statistics: Heart Disease. 1997. Available at www.Hawaii.gov/health/stats/vsheart.html. Accessed February 10,2003.
- 7.Motheral BR, Fairman KA. The use of claims databases for outcomes research: Rationale, challenge; and strategies. Clin Ther. 1997;19:346–66. doi: 10.1016/s0149-2918(97)80122-1. [DOI] [PubMed] [Google Scholar]
- 8.Clark DO, Von Korff MV, Saunders K, et al. A Chronic Disease Score with Empirically Derived Weight. Med Care. 1995;33:783–95. doi: 10.1097/00005650-199508000-00004. [DOI] [PubMed] [Google Scholar]
- 9.Fitzmaurice GM, Laird NM, Ware JH. Applied Longitudinal Analysis. John Wiley & Sons; 2004. [Google Scholar]
- 10.Greying JP, Denig P, van der Veen WJ, et al. Uptake of angiotensin II receptor blockers in the treatmer of hypertension. Eur J Clin Pharmacol. 2005;61:461–6. doi: 10.1007/s00228-005-0924-7. [DOI] [PubMed] [Google Scholar]
- 11.Psaty BM, Smith NL, Siscovick DS, et al. Health outcomes associated with antihypertensive therapie used as first-line agents. A systematic review and meta-anafysis. JAMA. 1997;277:739–45. [PubMed] [Google Scholar]
- 12.Psaty BM, Lumley T, Furberg CD, et al. Health outcomes associated with various antihypertensiv therapies used as first-line agents: A network meta-analysis. JAMA. 2003;289:2534–44. doi: 10.1001/jama.289.19.2534. [DOI] [PubMed] [Google Scholar]
- 13.ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group Major outcome in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calciur channel blocker vs. diuretic, the Antihypertensive and Lipid-Lowering Treatment to Prevent Heai Attack Trial (ALLHAT) JAMA. 2002;288:2981–2997. doi: 10.1001/jama.288.23.2981. [DOI] [PubMed] [Google Scholar]
- 14.Einhom PT, Davis BR, Massie BM, et al. The Antihypertensive and Lipid Lowering Treatment t prevent heart attack trial (ALLHAT) heart failure validation study: Diagnosis and prognosis. Am Hea J. 2007;153:42–53. doi: 10.1016/j.ahj.2006.10.012. [DOI] [PubMed] [Google Scholar]
- 15.Fischer MA, Avom J. Economic implications of evidence-based prescribing for hypertension. Ca better care cost less? JAMA. 2004;291:1850–1856. doi: 10.1001/jama.291.15.1850. [DOI] [PubMed] [Google Scholar]
- 16.Boltri JM, Gordon ER, Vogel RL. Effect of samples on physician prescribing patterns. Fam Med. 2002;34:729–731. [PubMed] [Google Scholar]
- 17.Giese M, Lackland D, Basile J, et al. 2003 Update on the hypertension initiative of South Caroline Bringing South Carolina from “worst to first” in cardiovascular health. JSC Med Assoc. 2003;99:157–161. [PubMed] [Google Scholar]
- 18.Siegel D, Lopez J, Meier J, et al. Academic detailing to improve prescribing patterns. Am J Hypertens. 2003;16:508–511. doi: 10.1016/s0895-7061(03)00060-8. [DOI] [PubMed] [Google Scholar]