RSS Feeds
Cognitive Behavioral Techniques for Hypertension: Are They Effective?
- David M. Eisenberg, MD;
- Thomas L. Delbanco, MD;
- Catherine S. Berkey, ScD;
- Ted J. Kaptchuk, BA;
- Bruce Kupelnick, BA;
- Jackie Kuhl, MA; and
- Thomas C. Chalmers, MD
- From Beth Israel Hospital, Harvard Medical School, and Harvard University School of Public Health, Boston, Massachusetts. Requests for Reprints: David Eisenberg, MD, Division of General Medicine and Primary Care, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215. Acknowledgments: The authors thank William Taylor, MD, and David Calkins, MD, for their scientific contributions and Deborah Arcarese for her technical and editorial assistance in the preparation of the manuscript. Grant Support: By a contract with the John E. Fetzer Institute and grants from the Waletzky Charitable Trust, the Henry J. Kaiser Family Foundation, and RO1 HS-05936 from the Agency for Health Care Policy and Research, U.S. Public Health Service, Department of Health and Human Services.
Abstract
Purpose: To assess by analysis of published controlled trials the efficacy of cognitive behavioral therapies (such as biofeedback, relaxation, meditation) for essential hypertension.
Data Identification: Randomized controlled trials published in the English language between 1970 and 1991 identified from the MEDLINE database and bibliographic references from these articles.
Study Selection: Limited to studies involving randomized assignment to a treatment group consisting of one or more cognitive behavioral interventions or a concurrent control group consisting of no therapy, a waiting list, regular monitoring, or placebo intervention.
Results of Data Synthesis: Although we identified more than 800 published works, only 26 met entry criteria. We identified a number of methodologic shortcomings, including small sample size, inconsistencies regarding baseline blood pressure determinations and types of control groups, and the possibility of confounding by multiple noncognitive cointerventions (diet, exercise) and expectancy (the placebo effect).
In meta-analyses involving 1264 patients, differences in mean blood pressure reduction varied according to the duration of baseline blood pressure measurements and type of control groups studied. In 16 comparisons involving baseline periods of more than 1 day, with patients (n = 368) assigned to either a cognitive therapy or a placebo intervention (sham biofeedback, pseudo-meditation), systolic and diastolic blood pressures decreased by 2.8 mm Hg (95% CI, 0.8 to 6.4) and 1.3 mm Hg (CI, 1.3 to 3.8), respectively. These changes were neither statistically nor clinically significant.
Conclusion: Cognitive interventions for essential hypertension are superior to no therapy but not superior to credible sham techniques or to self-monitoring alone. The literature on this subject is limited by a variety of methodologic inadequacies. No single cognitive behavioral technique appears to be more effective than any other.
Sixty million adult Americans are at increased health risk due to elevated blood pressure [1-3]. Essential hypertension was the most common diagnosis for patients visiting physicians' offices in 1990, occurring in 3.9% of visits [4]. Essential hypertension has been the most frequently reported morbidity-related diagnosis in every National Ambulatory Medical Care Survey since 1973 [4]. Several large, controlled clinical trials have established that lowering blood pressure reduces morbidity and mortality rate [5-7]. A meta-analysis of 14 randomized trials involving antihypertensive drugs (total, 37 000 persons; mean treatment duration, 5 years) concluded that a mean reduction in diastolic blood pressure of 5 to 6 mm Hg was associated with avoidance of at least one-third of the risk of stroke and at least one-fifth of the risk of coronary heart disease [8, 9]. However, these trials involved pharmacologic antihypertensive agents known to be associated with appreciable side effects [10-17] and had problems with compliance [18-20].
Data from controlled trials suggest that a number of behavioral interventions may be effective in lowering blood pressure. The literature is increasing regarding the application of noncognitive behavioral therapies, specifically diet and exercise, in the treatment of essential hypertension [21-41]. Our purpose was to assess the efficacy of cognitive behavioral therapies in the treatment of hypertension. Although there is no consensus as to which techniques are rightfully subsumed by the term behavioral medicine, meditation, relaxation, autogenic training, hypnosis, stress management, imagery, and biofeedback are among the cognitive interventions commonly included in attempts to define this area of investigation [42, 43].
We performed this analysis with the following study questions in mind: 1) What is the evidence that cognitive behavioral interventions are effective in the treatment of essential hypertension? 2) What is the estimate of the size of this effect, and how does it compare with estimates regarding pharmacotherapy for hypertension? 3) What are the principal scientific inadequacies of randomized, controlled trials in this area? and 4) How can existing methodologic problems be addressed to improve future clinical investigation in this area?
We present our results in two parts: a qualitative review of the strength of the evidence followed by a quantitative meta-analysis of published data from relevant randomized, controlled trials.
Methods
Selection of Papers
We used PaperChase (Beth Israel Hospital, Boston, Massachusetts), a program to search the MEDLINE database [44], to identify studies published in English between 1970 and 1991 involving cognitive behavioral techniques in the treatment of adult patients with mild to moderate hypertension (diastolic blood pressure of 90 to 114 mm Hg; systolic blood pressure > 140 mm Hg). At the time of our literature search, the data base contained references indexed through November 1991.
We conducted searches using the single medical subject (MeSH) heading, hypertension. By using the National Library of Medicine's designation of major topic, we restricted our searches to references in which this MeSH term was the major focus of the article. In addition, the following MeSH treatment terms were selected: autogenic training, behavior therapy, aversive therapy, biofeedback (psychology), cognitive therapy, desensitization (psychology), implosive therapy, relaxation techniques, hypnosis, suggestion, autosuggestion, stress, psychological and yoga. In addition, the following non-MeSH terms were selected: meditation, nonpharmaceutical, nondrug, and guided imagery.
References found by the treatment terms were pooled into a single list using the Boolean operator OR. The pooled list of treatment terms was then combined with the hypertension list using the Boolean operator AND.
For the purpose of this review and meta-analysis, articles needed to meet the following inclusion criteria: 1) randomization of adults with a history of mild to moderate essential hypertension (diastolic blood pressure, 90 to 114 mm Hg); 2) inclusion of at least one experimental group treated with a cognitive behavioral intervention; 3) a concurrent control group involving patients randomly assigned to either no therapy, a waiting list, regular monitoring, or a placebo intervention (sham therapy); 4) detailed descriptions of both experimental and control interventional techniques; and 5) detailed reporting of baseline blood pressures and blood pressures obtained after a predetermined treatment period. We excluded studies involving children, adults with secondary hypertension, nonrandomized studies, publications presenting follow-up data from earlier randomized, controlled trials, and studies limited to women with toxemia of pregnancy.
Drawing on the 857 articles identified by the literature search and references from these articles, we initially identified 36 trials for analysis [45-80]. Subsequent review by three investigators resulted in the rejection of 10 of these trials [71-80]. Three were found to be follow-up studies of controlled trials already accepted for analysis [71, 73, 79], five were judged to be nonrandomized trials [72, 74-77], and two did not contain extractable blood pressure data [78, 80]. The 26 trials ultimately selected for analysis were then cross-referenced with our initial MEDLINE search. Each of the 26 trials was present on our original computer-generated list of 857 references.
Identification of Cognitive Behavioral Techniques
The field of cognitive-behavioral therapies has not been well standardized. Therapies are often combined or overlap. Disagreement exists regarding their proper categorization and different texts categorize similar techniques differently. The MEDLINE literature search described above included terms that were popular in earlier periods (hypnosis, autosuggestion) and terms that are currently popular (guided imagery, stress management). Articles accepted into this study made explicit reference to the following cognitive behavioral therapies:
Biofeedback: A feeding back of physiologic activity with audio or visual devices to allow patients to gain conscious control of otherwise unconscious processes [81, 82].
Meditation: Techniques that focus attention and produce a self-aware state of inner calm [83, 84].
The Relaxation Response: Thought to be the opposite of the fight-or-flight response, invoked by repetition of a word or phrase in passive attitude with decreased muscle tone [85, 86].
Progressive Relaxation Techniques: Voluntary relaxation of selected muscle groups, often combined with calming imagery and focused breathing [87, 88].
Stress Management: Techniques that reduce excessive stress arousal by changing cognitive and emotional responses to events; these can include cognitive restructuring, adaptive emotional learning strategies, imagery, and physical or mental relaxation [89-92].
Qualitative Review
Sackett and coworkers [93] at McMaster University have suggested a number of specific criteria whereby studies can be analyzed methodologically to distinguish useful from useless and potentially harmful therapies. These criteria include the following:
1. Was the assignment of patients randomized?
2. Were all clinically relevant outcomes reported?
3. Were the study patients recognizable and similar to your own?
4. Were both clinical and statistical significance considered?
5. Was the therapeutic maneuver subject to contamination or cointervention?
6. Were all patients who entered the study accounted for at its conclusion, and was the drop-out rate acceptable?
We reviewed all prospective studies according to the six criteria mentioned above. Based on our qualitative review, we extracted quantitative data pertaining to a range of study variables (Table 1).
Blinding of Papers before Quantitative Analysis
A technician altered all papers so that readers were unaware of the source of the publication. Blinding of material was limited to the Methods and Results sections. All results were further blinded to refer to groups A, B, or C in the Results section. At least two investigators were asked to extract data for analysis from blinded materials. Discrepancies were discussed and, if necessary, adjudicated by a third investigator.
Quality Scoring of Individual Papers
Blinded papers were scored for quality of experimental design by a method previously described [94]. This method, applied only to randomized, controlled trials, develops a quantitative score for each paper that reflects both the reporting of procedures and the efforts to control biases that might influence the result. The score obtained is useful for ranking the papers according to quality. Two investigators scored the papers independently. Discrepancies were discussed with a third investigator, and a final score for each paper was established.
Data Analyses
The treatment effect of each experimental (cognitive) therapy was defined as the decline in blood pressure of the experimental (cognitive) group minus the decline in blood pressure of the control (waiting list or placebo) group. A positive treatment effect means that the experimental group experienced a larger blood pressure decline than the control group. Systolic and diastolic treatment effects were extracted from each study and refer to blood pressures obtained by a medical professional at baseline and after the initial treatment phase of each study. The data extracted from blinded papers were published summaries of baseline and post-treatment blood pressure levels.
Before we analyzed the data, we hypothesized that four factors would be related to therapeutic efficacy: 1) length of baseline period [ 1 day versus >1 day]; 2) type of control [waiting list or no therapy versus sham or placebo]; 3) scientific quality score; and 4) type of cognitive therapy. We performed a series of meta-analyses using the random effects model of DerSimonian and Laird [95] to provide summary estimates of treatment effect. Separate meta-analyses were performed on subgroups of studies defined by three of these factors: baseline period, type of control, and type of cognitive therapy. We report mean quality scores for each subgroup of studies. Although the heterogeneity observed among studies was reduced within subgroups (the meta-analyses were performed on more homogeneous subgroups of studies), the DerSimonian and Laird random-effects model corrected for any remaining heterogeneity. This method uses the treatment effects (cognitive therapy blood pressure reduction minus control reduction) from the individual studies and their standard errors to derive an overall estimate of treatment effect. In 11 studies where the standard errors were not directly reported, we relied on reported t-tests or confidence intervals to calculate treatment effect standard errors.
We present the findings of these meta-analyses on systolic and diastolic blood pressure data in terms of overall estimates of treatment effect and their respective 95% confidence intervals using the methods of DerSimonian and Laird [95].
Results
Qualitative Review
In accordance with intake criteria, all 26 studies met the first two McMaster criteria, namely that patients be randomly assigned to an experimental or control group and that outcome measures be reported in detail.
We observed a common inconsistency in how fully a patient's clinical status was described. Twelve studies reported baseline blood pressure determinations obtained over a single day of observation [46, 52-55, 58, 62, 66-70], whereas the remaining 14 studies reported baseline measurements obtained over a period ranging from several days to several months. Similarly with regard to blood pressure assessment after intervention, 17 studies [46, 50, 52, 53, 55-60, 62, 65-70] reported single-day blood pressure measurement or failed to comment on the time period used, whereas 9 studies [45, 47-49, 51, 54, 61, 63, 64] reported follow-up assessment obtained over more than 1 day. Furthermore, three studies [49, 51, 55] explicitly instructed patients to perform cognitive interventions at the time of their post-therapy assessment, whereas four studies [48, 61, 64, 69] explicitly told individuals not to perform cognitive interventions at the time of their assessment. Nineteen papers [45-47, 50, 52-54, 56-60, 62, 63, 65-68, 70] did not comment explicitly on the use or prohibition of cognitive interventions at the time of follow-up blood pressure assessment.
Experimental and control groups tended to have small sample sizes; one half of all groups had 13 or fewer persons. Jacob and colleagues [63] estimated that a sample size of 30 or more persons per group was required to detect a change in systolic blood pressure of 5 mm Hg at the 5% significance level with a statistical power of 0.80. Only 5 of 26 studies had sample sizes of this magnitude [54, 63, 65, 66, 68]. Ten studies [45, 51-55, 61, 62, 64, 67] did not address the issue of clinical significance in interpreting their results. Confounding or cointervention (the fifth McMaster criterion) was infrequently dealt with well. Roughly one in three studies [48, 51, 60, 61, 63, 65, 66, 70] raised the possibility that patients' expectancy of improvement in blood pressure control (nonspecific placebo effects) may be capable of confounding observations regarding changes in blood pressure attributable to cognitive interventions. Six studies involved credible sham interventions and involved some assessment of expectancy on the part of both experimental and control participants [46, 51, 57, 62, 64, 70].
Potential confounding by noncognitive behavioral factors (such as weight reduction, salt and alcohol consumption, and exercise) was reported to some extent in all 26 trials; however, only two studies presented detailed data on all these variables [63, 70]. Patients in 18 of 26 studies used concomitant oral antihypertensive medications at entry and continued to use these medications (sometimes with alterations in dosage) during the experimental or control periods of investigation [45-47, 49, 50, 52, 53, 55, 56, 58, 59, 61, 62, 64-67, 69]. Studies had little variability concerning the duration of cognitive intervention (mean, 9.7 weeks; SD = 3.7).
Drop-out rates (the sixth McMaster criterion) ranged from zero to 29.1%, with a mean of 10.0% for the 26 studies. One fourth of the studies had drop-out rates greater than 15.7%. In summary, our qualitative review uncovered a variety of methodologic inconsistencies and inadequacies.
Meta-Analyses (Quantitative Review)
For the entire study population, controls (n = 541) experienced mean absolute reductions in systolic blood pressure of 4.02 mm Hg (CI, 2.22 to 5.82) and reductions in diastolic blood pressure of 2.57 mm Hg (CI, 0.87 to 4.27).
Table 2 summarizes the mean reductions in systolic and diastolic blood pressures (estimated by the DerSimonian and Laird method [95]) among patients treated by each type of cognitive therapy (experimental groups versus control groups), while controlling for the type of control condition used and the duration of baseline assessment. We also report the number of trials and persons studied and mean quality scores for the trials within each group.
The cognitive intervention group titled relaxation techniques includes studies with one or more experimental groups assigned to the regular elicitation of the relaxation response [85, 86] or progressive relaxation techniques [87, 88].
Figures 1 and 2 graphically depict global findings from Table 2. Differences in the mean reductions in blood pressure were highest when comparing experimentally treated individuals with those who received no therapy and for whom baseline blood pressure assessments were made during a single day. Under these conditions, patients treated with cognitive behavioral therapies experienced a mean reduction in systolic blood pressure of 13.4 mm Hg (CI, 9.3 to 17.6 mm Hg) and reductions in diastolic blood pressure of 9.0 mm Hg (CI, 5.2 to 12.9 mm Hg) beyond those of controls (P < 0.05). Mean blood pressure reductions were smallest when comparing persons experimentally treated with those randomly assigned to a credible placebo or sham intervention and for whom baseline blood pressure assessments were made during a period of more than 1 day. Under these conditions, individuals treated with cognitive behavioral therapy experienced a mean reduction in systolic blood pressure of 2.8 mm Hg (CI, 0.8 to 6.4 mm Hg) and a mean reduction in diastolic blood pressure of 1.3 mm Hg (CI, 1.3 to 3.8 mm Hg) relative to controls. Reductions in blood pressure under these latter conditions were not statistically significant (P > 0.1).
For studies involving sham or placebo control groups with baseline periods of more than one day (Table 2), no single cognitive intervention was found to be statistically superior to any other in reducing blood pressure. The data suggest that overall reductions in systolic blood pressure were higher in studies in which patients received multiple cognitive therapies compared with studies in which patients received a single cognitive intervention (P < 0.05).
In the one study that included groups randomly assigned to either relaxation, biofeedback, antihypertensive medication, or a credible placebo control (mild exercise) [57], patients receiving antihypertensive medications (n = 10) had mean reductions in systolic and diastolic blood pressure of 13.1 mm Hg and 5.8 mm Hg, respectively, beyond those of the placebo control group (n = 11). These reductions relative to placebo were statistically greater than reductions observed among patients who received relaxation therapy (reductions of 6.5 mm Hg, systolic; and 1.0 mm Hg, diastolic) or biofeedback (reductions of 2.2 mm Hg, systolic and 2.9 mm Hg, diastolic).
Discussion
In recent decades the medical community has sought to measure the effects of cognitive function on physiology through the application of technology and clinical experimentation. From the 1940s through the 1960s, investigators used electrocardiogram machines to document the cardiovascular changes induced by Indian yogis during meditation [96-98]. In the 1970s Wallace and colleagues [99, 100] documented acute physiologic changes in advanced practitioners of transcendental meditation. Subsequent experiments extended these findings and confirmed the human capacity to alter physiology at will, relying solely on cognition [101, 102].
During the past two decades, proponents of behavioral medicine have advocated that cognitive behavioral interventions be applied to selected patients with the intent of altering the natural course of disease. This strategy is based on the supposition that cognitive behavioral techniques, known to alter physiology acutely under laboratory conditions in highly trained individuals (advanced meditators, for example), can induce chronic physiologic changes when taught to, and practiced by selected patient groups. Furthermore, it has been hypothesized that these chronic physiologic changes can predictably alter the course of common diseases.
Our literature search yielded more than 800 works relevant to the application of cognitive behavioral techniques to adults with essential hypertension, but only 26 were randomized, controlled trials with data conducive to formal meta-analysis. These 26 trials involved a multitude of cognitive behavioral interventions (and combinations thereof) compared with the application of various control conditions. This literature is heterogeneous and there are fewer randomized, controlled trials than we had anticipated.
We performed a subsequent search of available psychology data bases to explore the possibility that our MEDLINE search failed to identify articles for analysis. This search identified 175 titles unique to the psychology data base; however, only two works met entry criteria for this analysis. The study by Jorgensen and colleagues [52] reported that hypertensive persons treated by anxiety management training showed decreases in systolic and diastolic blood pressure as compared with individuals in a waiting list control group. In the study by Hoelscher and colleagues [103], hypertensive adults were assigned to one of four relaxation conditions or to a waiting list control group. All relaxation conditions showed significantly greater reductions in systolic and diastolic blood pressures than waiting list controls but did not differ from each other. If data from these two studies were added to our meta-analysis (Figures 1 and 2), they would not change the statistical significance or overall conclusions of our analysis.
A limitation of this analysis arises from patients' heterogeneity and our pooling of data without stratification by sociodemographic or disease-specific characteristics. Few papers reported data regarding specific characteristics of individuals (age, race, gender, medication use) relative to baseline blood pressure or changes in blood pressure following cognitive intervention. Although all of the people in this analysis had essential hypertension, existing data are insufficient for analyses of subsets of patients treated with individual cognitive behavioral therapies. As such, our meta-analysis, which uses a random effects model, should be viewed as a rough and ready approach suitable for pooling heterogeneous studies.
Concerning overall technical quality, the 26 trials in this review had a mean quality score of 34 (maximum possible score, 104). This compares poorly with the technical quality scores of 376 randomized, controlled trials in 18 other therapeutic areas studied previously but is better than those performed in several fields [104, 105].
Our meta-analyses show that changes in blood pressure among patients receiving cognitive therapy are influenced by study design. Specifically, patients in studies using baseline periods of 1 day showed reductions in blood pressure that were far greater relative to controls than those observed in studies with baseline periods of more than 1 day. We cannot explain this phenomenon, but it may represent different rates of regression to the mean or acclimatization of individuals over time to the research environment, with concomitant reductions in blood pressure [106, 107]. Engel and colleagues [108] have reported that both systolic and diastolic blood pressures decrease predictably throughout the first 21 days of routine baseline measurement.
Options regarding the calculation of preintervention baseline blood pressures include 1) measurements obtained during a single day [not ideal given data from this analysis]; 2) an average of multiple blood pressures obtained during the entirety of a predetermined baseline period; 3) or some average of baseline readings taken during the final portion a predetermined baseline period, thereby diminishing presumed effects of acclimatization. In most published trials involving pharmacologic therapy for mild to moderate hypertension, estimates of baseline blood pressure were based on either a single reading or the average of two consecutive readings obtained during a single baseline visit [8, 9]. The 26 studies reviewed here have dissimilar definitions of baseline blood pressure assessment. Similar inconsistencies exist with regard to the determination of blood pressure following intervention. Without consensus, comparisons of baseline blood pressure data and differences in blood pressure reductions attributable to cognitive (or noncognitive) interventions will be impossible to interpret.
Another methodologic problem is the type of control condition used in assessing the relative efficacy of cognitive therapies. Our meta-analysis documented that hypertensive persons who received cognitive therapy showed statistically significant reductions in blood pressure relative to individuals receiving no therapy or placed on a waiting list. These reductions in blood pressure were clinically significant (reductions in both diastolic and systolic blood pressure were greater than 5 mm Hg, P < 0.05) for studies involving a 1-day baseline assessment. Hypertensive persons treated with cognitive therapies, however, showed reductions in blood pressure roughly half those noted above relative to individuals treated with a credible placebo or sham intervention. In this latter circumstance, among studies involving baseline periods of more than 1 day, blood pressure reductions were 2.8 mm Hg in systolic blood pressure and 1.3 mm Hg in diastolic blood pressure. These changes were neither statistically nor clinically significant.
Cognitive behavioral therapies prescribed in the absence of other behavioral interventions (such as weight loss, salt and alcohol reduction, increased exercise) are not nearly as effective as standard antihypertensive pharmacotherapy. Specifically, MacMahon and colleagues [8] performed a meta-analysis on 14 unconfounded randomized trials of antihypertensive drugs (chiefly diuretics or -blockers) with baseline diastolic blood pressure based on either a single reading or the average of two consecutive readings at a single baseline visit. Subsequent analyses revealed a mean reduction in diastolic blood pressure of 5 to 6 mm Hg and a mean reduction in systolic blood pressure of approximately 10 to 12 mm Hg relative to control group (placebo) reductions [9]. Data from our meta-analysis suggest that cognitive interventions alone are not capable of blood pressure reductions of this magnitude.
Our inability to document the efficacy of cognitive therapies in the treatment of essential hypertension under the above-mentioned controlled conditions (relative to a credible sham intervention) is consistent with the Canadian Consensus Report on Non-Pharmacological Approaches to the Management of High Blood Pressure [109] and a recent report of the Trials of Hypertension Prevention Collaborative Research Group [23].
Only 1 of the 26 studies [59] documented changes in health care costs associated with cognitive therapeutic management of individuals with essential hypertension. An article by Glasgow and colleagues [110], involving a stepped care approach to the management of hypertension, suggests that a program involving self-monitoring and cognitive behavioral intervention can reduce medication use and overall health care costs. Clearly, more data are required before definitive costbenefit analyses can be performed.
A principal confounder in this literature is that of expectancy of improvement or nonspecific therapist-trainee interactions (the placebo effect). More than a generation ago Beecher warned that . There is too little scientific as well as clinical appreciation of how important unawareness of these placebo effects can be and how devastating to the experimental studies as well as to sound clinical judgement lack of attention to them can be [111]. Our analyses suggest that patients' expectancy of clinical improvement is, indeed, a potential confounder of blood pressure modification and that a positive expectation, when coupled with a credible placebo or sham intervention, can, in some instances, result in statistically (and clinically) significant reductions in systolic and diastolic blood pressures. This observation highlights the contradictory nature of the so-called placebo effect. On one hand, a failure to control for expectancy may result in observed increases in the difference in clinical outcomes between experimental and control groups. On the other hand, it appears from existing data involving studies in which patients received either cognitive therapy or no therapy that certain patients under certain circumstances can be taught to harness expectations that induce reductions in blood pressure comparable to those attributed to standard pharmacotherapy (diastolic blood pressure reduction 5 mm Hg). We lack the knowledge to predict which individuals, under which circumstances, are likely to manifest these physiologic changes. Moreover, we do not yet possess a neurochemical explanation as to how this complex psychophysiologic dialogue is conducted.
Future clinical research to assess the efficacy and mechanisms of action of cognitive behavioral interventions for hypertension should, ideally, incorporate the following methodologic strategies:
1. Study participants should be recruited from relatively large demographic cohorts (e.g., health maintenance organizations, multiple corporate worksites) to ensure adequate sample size and generalizability to the adult U.S. population.
2. Designs must address the question of whether it is ethical to assign any patient group to a no-therapy condition.
3. Studies should involve nonmedicated individuals or those for whom antihypertensive medications are prescribed according to a strict algorithm applied equally to all groups.
4. Cognitive interventions should be standardized and sufficiently described to be implemented by a variety of providers.
5. Interventions should take place at multiple sites in an effort to address potential confounding by nonspecific therapist-patient interactions.
6. Individuals' expectancy of improvement should be assessed by questionnaire before random assignment, during the experimental intervention, and again at the end of the intervention period (but before results are known).
7. Reasons for exclusion of individuals or their subsequent drop-out from the study should be reported in detail.
8. Baseline assessment should last at least 3 weeks and authors must specify how baseline calculations are performed.
9. Efforts should be made to standardize post-treatment blood pressure determinations and methods whereby blood pressure differences (baseline versus post-treatment assessment) are calculated.
10. Outcome variables should include an assessment of side effects of treatment (side effects of drug or nondrug interventions) and an estimate of costs of cognitive or drug therapies, in addition to alterations in blood pressure and associated reductions in cardiovascular risk.
In light of the high prevalence of essential hypertension and the equivocal efficacy of cognitive behavioral treatments for this condition, further research in this area is needed. Future studies should ideally compare the efficacy of cognitive behavioral interventions with noncognitive behavioral therapies (exercise, diet), or drug therapies, or combinations of behavioral and drug therapies. Continued use of a no-therapy control group may not be ethically justifiable in light of existing information regarding cardiovascular risks associated with untreated hypertension.
- Copyright 2004 by the American College of Physicians
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
Related Clinical Content
