Patient and Provider Perceptions of Hypertension Treatment: Do They Agree?

Abstract

The objective of our study was to explore physician and patient attitudes regarding hypertension management. One hundred forty‐five primary care providers and 189 patients with hypertension at 6 Veterans Administration clinics completed a hypertension survey. Fifty‐one percent of patients were at their blood pressure goal, 58% were on guideline‐concordant therapy, and 31% achieved both. Patients and providers agreed that physicians were a “very/extremely” useful source of information but differed in perceived value of pharmacists, educational material, advertising, and the Internet. They also agreed on the value of preventing cardiovascular events but differed in their perceptions of the importance of medication costs, side effects, and national guidelines. Blood pressure control and guideline‐concordant therapy was higher than most prior reports, but with opportunity for improvement. Patients and providers differed in perceived value of various aspects of hypertension management; this information may help to determine trial design and quality improvement strategies in the future.

Hypertension affects an estimated 50 to 60 million Americans and 1 million veterans in the Veterans Administration (VA) health care system. ¹ , ² Although evidence‐based guidelines have been available for decades from the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC), identification and treatment of patients with hypertension remains suboptimal. The seventh report of the JNC (JNC 7) and VA guidelines suggest that thiazide diuretics should be first‐step therapy for uncomplicated hypertension and should be added more frequently to current regimens to improve blood pressure (BP) control. ³ , ⁴ Recommendations are based on the results of the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) and many other trials. ⁵

In spite of knowledge about hypertension treatment results, there is a “quality gap” between evidence based guidelines and patients on appropriate therapy who are at goal BP. ⁶ , ⁷ A recent VA study found that only 43% of hypertensive patients attained BP <140/90 mm Hg and only 60% of patients were on therapy consistent with present JNC 7 guidelines (“JNC 7 guideline‐concordant therapy”). ⁸ In a study of 20 private‐sector clinics, only 50% of hypertensive patients had a BP <140/90 mm Hg and only 35% were on JNC 7 guideline‐concordant therapy ⁹ ; the 1990–2000 National Health and Nutrition Examination Survey (NHANES) found that only 31% of hypertensive patients had BP <140/90 mm Hg. ¹ Several studies have shown that despite familiarity and agreement with national hypertension guidelines, providers often tolerate BP higher than recommended. ⁷ , ⁸ To close this gap, numerous studies involving academic detailing, ¹⁰ , ¹¹ promotion of guideline adherence, ¹² , ¹³ and systems‐based interventions such as computerized reminders have evaluated physicianbased interventions. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ All of the studies attempted to overcome provider and health care system limitations, with varying degrees of success.

The objective of our study was to survey patients and providers in 2 regional VA health care systems to better understand physician and patient attitudes regarding hypertension management and to identify barriers and opportunities for guidance in designing clinical trials and quality improvement initiatives for hypertension care.

METHODS

Participants

Primary care providers and patients were recruited from the Iowa City and Minneapolis VA Medical Centers (VAMCs) and 4 affiliated communitybased outpatient clinics (CBOCs) from September 2004 to November 2004. Eligible providers included all 243 physicians, residents, nurse practitioners (NPs), and physician assistants (Pas) who provided primary care at the 6 clinics. Providers were recruited via e‐mail and completed anonymous questionnaires online.

Eligible patients included a convenience sample of 191 primary care patients with hypertension waiting to be seen by their providers and asked to participate in a 15‐minute survey on BP treatment. A structured interview was completed in person by a member of the research team. Patients were offered a $5 phone card and providers were offered a $10 VA canteen gift certificate for completing the survey.

Data

The first section of the provider survey was a 15‐question multiple‐choice knowledge quiz designed to assess familiarity with JNC 7 guidelines previously developed by one of the investigators (BLC) with correct answers based on JNC 7 guidelines. ¹⁸ Brief clinical vignettes, previously used by Ubel and colleagues, ¹⁹ were used and providers were asked about their experience with angiotensin‐converting enzyme (ACE) inhibitors, β‐blockers, calcium channel blockers, and thiazide diuretics in uncomplicated hypertension (defined as hypertension without comorbid diabetes mellitus, chronic heart failure, coronary artery disease, or renal insufficiency) with regard to effectiveness and tolerance. Providers were then asked to rate the importance of numerous factors influencing their choice of antihypertensive medications on a 5‐point Likert scale, ranging from “not at all important” to “extremely important.” The same scale was used to rate how important they believed those same factors were to their patients. They were asked to rate how useful they believed several sources of information were to their patients including the following: their doctor, their pharmacist, educational material from the VAMC, advertising such as TV or magazines, and the Internet. A similar 5‐point Likert scale was used for responses, ranging from “not at all useful” to “extremely useful.” The 5‐point scale was dichotomized with responses 1 to 3 in one group and 4 to 5 in another to simplify the results. Results were similar when means were compared, but only dichotomized percentages are reported. Complete surveys are available at: http://www1. va.gov/CRIISP/research_tools.htm.

The patient survey included demographics and the importance of the factors listed above on the same 5‐point scales used by providers. Additional information was collected from the VA's Computerized Patient Record System (CPRS) including current hypertension medications and intolerances, BP in the past 12 months, comorbid illnesses, whether patients were at their BP goal, and whether they were on JNC 7 guideline‐concordant therapy. BP (BP) goal was defined as <140/90 mm Hg, or <130/80 mm Hg for patients with diabetes. Two physician investigators (PJK, AI) reviewed each patient chart and listed what changes might be made to achieve goal BP and/or JNC 7 guideline—concordant therapy. The physician investigators reviewed the JNC 7 guidelines, discussed the process of implicit chart review, and compared interpretation of guidelines. No formal interrater reliability was assessed, however.

The institutional review boards of the University of Iowa Carver College of Medicine, Iowa City VA Health Care System, and Minneapolis VAMC approved the study and all patients provided written informed consent. Providers were given a cover letter explaining the risks and benefits of the survey and that participation was voluntary.

Analysis

Descriptive statistics were generated for both samples. Differences between groups were analyzed using student t tests and analyses of variance for continuous variables and chi‐square tests for categoric variables. P values <.05 were considered significant. All analyses were performed using SAS software version 8.0 (SAS Institute Inc, Cary, NC).

RESULTS

Participant Characteristics

Of 191 patients asked to participate in the survey, 189 agreed (99%). Descriptive characteristics of the patient sample are included in Table I; the mean age was 66 years, 97% were men, and 92% were white. For the prior year, the mean BP for the entire sample was 140/76 mm Hg. Fifty one percent of patients were at BP goal, 58% were on JNC 7 guideline‐concordant therapy, and 31% were both at goal and on JNC 7 guideline‐concordant therapy. The most common comorbid illnesses were hyperlipidemia, diabetes, and coronary artery disease.

Table I.

Characteristics of 189 Patients with Hypertension From 6 VA Clinics in Iowa and Minnesota

Age, mean ± SD	65.8±11.0
Men, %	96.8
Race, %
White	92.0
Black	6.4
Hispanic	0.5
Other	1.1
Education
High school graduate or less, %	45.7
Some college or college graduate, %	54.3
Mean systolic BP prior year ± SD	140±22.5
Mean diastolic BP prior year ± SD	76±10.5
Mean BP recordings in prior year ± SD, No.	4.0±2.3
At BP goal, %	50.8
On JNC 7 guideline‐concordant therapy, %	58.2
Both at BP goal and on JNC 7 guideline‐concordant therapy	31.2
Top 10 Comorbid Illnesses, %
Hyperlipidemia	55.0
Diabetes	36.0
Coronary artery disease	25.4
Benign prostatic hypertrophy	19.0
Depression	16.9
COPD/asthma	15.9
Renal insufficiency	13.8
Smoking	12.7
Sleep apnea	9.0
Atrial fibrillation	7.4
Abbreviations: VA, Veterans Administration; BP, blood pressure; JNC 7, Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; COPD, chronic obstructive pulmonary disease.

Of the 243 primary care providers at the 6 facilities, 145 completed the online survey (60%). Of note, 62% were residents, 29% were staff physicians, and 9% were Pas or NPs.BP medication use is listed in Table II, with ACE inhibitors most frequently used overall (55%) and most frequently used as monotherapy (43.9%). Thiazides were the fourth most frequently used overall (33.3%) and third most frequently used as monotherapy (14.6%).

Table II.

Blood Pressure Medication Use and Intolerance in 189 Patients with Hypertension

Medication Class	Patients Using (N=189)	Patients Using as Monotherapy (N=41)	Patients with Intolerance (N=189)
ACE inhibitor	104 (55.0)	18 (43.9)	12 (6.3)
β‐Blocker	99 (52.4)	11 (26.8)	10 (5.3)
Calcium channel blocker	66 (34.9)	4 (9.8)	8 (4.2)
Thiazide	63 (33.3)	6 (14.6)	8 (4.2)
Loop diuretic	51 (27.0)	0 (0)	2 (1.1)
α‐Blocker	29 (15.3)	0 (0)	4 (2.2)
Nitrates	23 (12.2)	0 (0)	0 (0)
Other (eg, hydralazine, clonidine)	18 (9.5)	1 (2.4)	5 (2.6)
ARB	12 (6.4)	1 (2.4)	0 (0)
Values are expressed as No. (%). Abbreviations: ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker.

Table III displays provider perceptions of what sources of information are very/extremely useful to patients in managing their BP and compares them with the actual patient responses. Both patients and providers valued the importance of the patient's physician in providing very/extremely useful information (77.9% vs 79.6%; P=.66). Providers significantly overestimated the importance of pharmacists (68.7% vs 56.7%; P=.02), advertising (10.5% vs 3.8%; P=.01), and the Internet (16.7% vs 6.3%; P<.01), however, while significantly underestimating the importance of VA educational material to patients (27.8% vs 40.7%; P=.01).

Table III.

Sources of Hypertension Information Deemed “Very/Extremely Useful” to Patients

	Provider Perception to Patient of“Very/Extremely Useful” Information Sources (N=145)	Patient Opinion of “Very/Extremely Useful” Information Sources (N=189)	P
Physician, %	77.9	79.6	.66
Pharmacist, %	68.7	56.7	.02
Educational material from Veterans Administration, %	27.8	40.7	.01
Advertising (eg, television or magazines), %	10.5	3.8	.01
Internet, %	16.7	6.3	<.01

Table IV presents a comparison of factors rated by both physicians and patients as very/extremely important for the prescribing of antihypertensive medications. Of note, providers are significantly less concerned about the cost of a medication to the patient than are patients (51.2% vs 64.6%;P=.01), while their perception of the importance to patients was significantly overestimated (75.7% vs 64.6%; P=.03). There was also a highly significant difference between patients and providers with regard to the importance of antihypertensive medication cost to the VA. Only 16.8% of providers cited this as a very/extremely important factor in their prescribing practices, and even fewer (9.7%) perceived cost to the VA as very/extremely important to their patients; however, 61.9% of patients cited this as very/extremely important (P<.01 for both). Providers were less concerned about side effects of antihypertensive medications than patients (70.8% vs 84.1%; P<.01), and were less likely to believe that the effectiveness of the medication to prevent strokes and heart attacks was important to the patient (62.3% vs 95.2%; P<.01). Finally, a majority of both providers and patients felt that national guidelines recommending a medication were very/extremely important (70.8% vs 67.7%; P=.55); however, only 21.7% of providers thought it was important to patients (P<.01).

Table IV.

Factors Related to Antihypertensive Prescribing Reported as “Very/Extremely Important” to Provider, Provider Perception of Importance to Patient, and Patient Opinion

	Provider Opinion (N=145)	Provider Perception of Importance to Patient (N=145)	Patient Opinion (N=189)	P (Provider vs Patient)	P (Provider Perception vs Patient)
Cost of medication to patient, %	51.2	75.7	64.6	.01	.03
Cost of medication to Veterans Administration, %	16.8	9.7	61.9	<.01	<.01
Adverse effects, %	70.8	81.2	84.1	<.01	.47
Preventing stroke or heart attack, %	95.7	62.3	95.2	.87	<.01
National guidelines recommending a medication, %	70.8	21.7	67.7	.55	<.01

When asking patients about the importance of the cost of their medications, 84% of patients reported that the reduced price of prescriptions at the VA was very/extremely important to them and 27% reported that it was the only reason or the main reason they came to the VAMC for care. Ninety percent of patients reported they filled more than 75% of their medications at the VA pharmacy and 56% paid a copay for their medications. Of those with a copay, 40% cited that this was a significant financial burden to them, but only 20% reported not filling a prescription because of cost.

Provider Knowledge

When asked about JNC 7 guidelines, 96% of providers reported having “heard of them,” 63% were “very familiar” with the report, and only 68% agreed with the recommendation of thiazides as first‐step therapy for uncomplicated hypertension. On the 15‐item quiz, VA providers scored 66% correct. These correct responses were consistent with previous findings in which family medicine residents scored 58% correct; primary care faculty, 72% correct; and certified hypertension specialists, 85%. ¹⁸ Pas and NPs scored significantly lower than staff and resident physicians, however (Table V).

Table V.

Provider Knowledge Survey Results, Stratified by Provider Type

	All Providers (N=145)	Staff (n=42)	Resident (n=90)	PA/NP (n=13)	P
JNC 7 knowledge quiz, % correct	65.7–16.3	68.3–16.4	66.8–13.7	50.0–23.8	<.01
“Which class of medication would you prescribe as initial monotherapy for a patient with uncomplicated hypertension?”
ACE inhibitor	9.4	9.5	8.9	7.7	.98
β‐Blocker	15.9	11.9	16.7	23.1	.59
CCB	0.7	2.4	0.0	0.0	.29
Thiazide diuretic	68.3	64.3	72.2	53.9	.33
“In your experience, what percent of the time will each of the following drugs (as monotherapy) achieve normal blood pressure?”
ACE inhibitor	39.3	45.5	36.7	39.5	.12
β‐Blocker	39.8	41.8	38.4	44.5	.55
CCB	37.5	44.4	35.0	35.5	.08
Thiazide diuretic	36.9	41.2	34.5	42.3	.21
“In your experience, what percent of the time do patients have to discontinue each of the following due to adverse effects?”
ACE inhibitor	15.1	14.0	15.0	19.5	.23
β‐Blocker	15.3	15.0	15.0	18.6	.65
CCB	12.7	10.0	13.2	16.8	.11
Thiazide diuretic	10.9	7.8	11.7	14.0	.14
“Which of the following classes of drugs have been proven to reduce the risk of stroke in hypertensive patients?” (check all applicable.)
ACE inhibitor	50.0	55	50	38	.59
β‐Blocker	71.0	64	74	69	.49
CCB	24.1	31	20	31	.33
Thiazide diuretic	53.1	81	40	54	<.001
“Which of the following classes of drugs have been proven to reduce mortality in hypertensive patients?” (check all applicable.)
ACE inhibitor	65	55	71	54	.13
β‐Blocker	78	71	80	85	.46
CCB	17	14	17	23	.76
Thiazide diuretic	53	81	41	46	<.001
Values are expressed as percentages. Abbreviations: PA, physician assistant; NP, nurse practitioner; JNC 7, Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; ACE, angiotensin‐converting enzyme; CCB, calcium channel blocker.

The final set of results describes providers' attitudes toward classes of antihypertensive medications. When asked “Which class of medication would you prescribe as initial monotherapy for a patient with uncomplicated hypertension?”, thiazides were selected by 68% of respondents, followed by 16% for β‐blockers, with no difference by provider type. For the second question, “In your experience, what percent of the time will each of the following drugs (as monotherapy) achieve a normal BP?”, all 4 agents were considered equally effective with 37% to 40% of patients achieving normal BP with monotherapy, with no difference by provider type. The third question was, “In your experience, what percent of the time do patients have to discontinue each of the following due to adverse effects?” Discontinuation rates ranged from 11% for thiazides to 15% for ACE inhibitors, with no difference by provider type.

Efficacy of antihypertensive agents was assessed by asking, “Which of the following classes of drugs have been proven to reduce the risk of stroke in hypertensive patients”? β‐Blockers were considered most effective and calcium channel blockers, the least. The only difference across providers was that staff physicians were twice as likely to believe that thiazides reduce the risk of stroke compared with residents (81% vs 40%; P<.001). Results were similar when asked “Which of the following classes of drugs have been proven to reduce mortality in hypertensive patients”?

COMMENT

The overall purpose of this survey was to better understand patient and provider attitudes toward the management of hypertension. Some important themes emerged that might help to guide the development of future research and quality improvement initiatives. The first finding was that only 50% of patients in this sample were at BP goal, 58% were on JNC 7 guideline‐concordant therapy, and only 31% were at both. These rates are similar to a recent VA study by Steinman and associates ⁸ and are higher than most private sector rates. ¹ , ⁹ Although a 100% rate is not attainable due to changes in BP and an ever‐changing prevalence, our results indicate that there is still room for improvement.

Another important finding was the discordance between physician knowledge and belief and actual practice. Even though 68% of providers agreed with JNC 7 guidelines that thiazide diuretics should be considered first‐step therapy for uncomplicated hypertension and 68% of providers would chose a thiazide for initial monotherapy, only 33% of patients were actually prescribed a thiazide. This gap between belief and actual practice and is not well understood but is consistent with recent work by one of the authors (BLC) that showed no relationship between physician hypertension knowledge and their patient's BP control. ¹⁸ These findings support studies that have shown minimal efficacy of physician‐targeted education to improve hypertension control. It also suggests that the level of clinical inertia is high and interventions should target this aspect of the health care delivery.

A positive finding is that acceptance of thiazides as initial monotherapy has improved over time. In a prior study by Ubel and colleagues ¹⁹ using the same question, researchers found that only 18% of physicians recommended thiazides as initial monotherapy for hypertension, whereas we report a rate of 68%, suggesting that the message from the JNC 7 and ALLHAT may be influencing providers. The authors also reported that 22% of providers thought that β‐blockers had the highest rate of adverse effects leading to discontinuation, whereas we report a 15% rate. This finding suggests that over time, providers may be less concerned about the adverse effect profile of β‐blockers.

We also found differences between patients' and providers' perceptions on where patients receive very/extremely useful information on hypertension. Physicians overestimated the importance of pharmacists, advertising, and the Internet and underestimated the importance of educational material from the VA. These findings underscore the potential benefit of educational material from the VA to patients but also show how little value this patient sample places on advertising and the Internet. With only 6.3% of respondents believing the Internet to be very/extremely useful, it remains to be seen how useful veterans will find the new Web‐based resource from the VA, MyHealtheVet. ²⁰

The perception of the cost of medications was also an important finding. Only half of providers felt that the cost of medications to the patient was very/extremely important and overestimated the perceived importance to the patient, relative to the patient's opinion. This finding does suggest a high level of sensitivity of medication cost burden by physicians for their patients. Alternatively, 62% of patients felt that the cost to the VA was very/extremely important to patients, whereas only 17% of providers reported the cost to the VA was very/extremely important to providers. This finding suggests that veterans are both cost‐sensitive as it relates to themselves, as well as cost to the VA and society. This may be a unique finding to veterans as they may feel a higher level of social responsibility to VA health care, but it demonstrates a lack of concern for cost to the VA on the part of providers. This has important ramifications for implementing cost‐containment strategies in VA and other health care systems. ²¹

Physicians and patients agreed on the importance of preventing strokes and heart attacks (95% felt it was very/extremely important), yet only 62% of physicians perceived that it was very/extremely important to patients. We have no explanation as to why 38% of physicians felt that reducing strokes and heart attacks was not very/extremely important to patients, since this is the primary reason for treating hypertension. The finding that 71% of physicians and 68% of patients felt that national guidelines recommending a medication was very/extremely important suggests that stressing guidelines to providers and patients is one potential component of interventions to improve hypertension care. This information suggests that national guidelines should be disseminated to patients and the public in addition to providers.

Another aspect of provider knowledge was the difference in scores by provider type. Overall, Pas and NPs scored significantly lower on a JNC 7 knowledge quiz, with only 50% correct compared with 67% for residents and 68% for staff physicians (P<.01). This is consistent with the finding that only 54% of Pas and NPs recommended thiazide diuretics for initial monotherapy, compared with 72% of residents and 64% of staff physicians. These findings may represent differences in level of training or simply reflect the small sample size of Pas and NPs (n=13).

One inconsistent finding is that resident physicians were less likely to believe that thiazide diuretics reduced the risk of stroke (40% vs 81% for staff; P<.001) or agree that thiazides reduce mortality (41% vs 81% for staff; P<.001), in spite of similar rates of recommendations for initial monotherapy, achievement of normal BP, or rates of adverse effects for thiazides. This inconsistency in knowledge suggests that provider education is only one component of clinical interventions to improve care. Because knowledge alone does not improve care, interventions should target multiple aspects of hypertension care. A recently published study of patient and provider intervention found the greatest improvement in care with a targeted patient component in addition to provider education and computer alerts. ²²

Our findings are consistent with a recently published national Harris Poll survey of 1245 patients with hypertension. ²³ Our samples had similar rates of obesity, hyperlipidemia, and diabetes, although our population was primarily white. In the national survey, patient self‐reported control rate (at the last BP taken) was 60% compared with the 51% rate we found through chart review, and overall thiazide use was 56% compared with our rate of 33% (this included thiazide combination therapy in the Harris Poll survey). In both surveys, patients placed a high value on information from their provider, but only 6% of patients received hypertension information from the Internet.

There are a number of limitations that should be mentioned. First, generalizability beyond a mostly white, male, midwestern, VA patient population and VA providers is limited. We included both urban and rural populations and our response rate was almost 100%, however. Our patient sample was a convenience sample of consecutive patients in primary care clinics, which is not a true random sample. We sampled all primary care providers and achieved a 60% response rate, which is a generally accepted response rate for physicians ²⁴ but has the potential for response bias. Both patient and provider surveys may suffer from social desirability response bias. ²⁵ Patients were not matched with their physicians, so actual prescribing information may differ from the physician survey. Because of a lack of standards for determining JNC 7 guideline‐concordant therapy and interrater reliability testing between the 2 physician reviewers, we may not have accurately determined JNC 7 guideline concordance.

CONCLUSIONS

Our findings suggest that the quality of hypertension care remains suboptimal but may be improving. Educational and quality improvement initiatives should be multifaceted to target patient, provider, and system factors that contribute to hypertension care. The use of thiazide diuretics needs to be promoted to both patients and providers, especially to counter the marketing influence of the pharmaceutical industry on both. ²⁶ Finally, the discordance between patient and provider perceptions of important factors in managing hypertension needs to be addressed to optimize patient‐provider relationships and improve clinical outcomes.

Disclosures: This work was presented at the 2006 VA Health Services Research and Development Annual Meeting, Arlington, VA, and the 2006 Society of General Internal Medicine Annual Meeting, Los Angeles, CA. The work reported here was supported by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service (IMV 04–066–1 and HFP 04–149). Complete surveys available at: http://www1.va.gov/CRIISP/research_tools.htm. Dr Kaboli is supported by a Research Career Development Award from the Health Services Research and Development Service, Department of Veterans Affairs (RCD 03–033–1). Dr Carter is supported by the National Heart, Lung, and Blood Institute, R01 HL069801 and R01 HL070740, and the Center for Research in the Implementation of Innovative Strategies in Practice (CRIISP) at the Iowa City VA Healthcare System, Iowa City, IA. Dr Ishani is supported by the Center for Epidemiology and Clinical Research, Minneapolis Veterans Administration Medical Center, Minneapolis, MN. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.