A Simplified Approach to the Treatment of Uncomplicated Hypertension
A Cluster Randomized, Controlled Trial
Notwithstanding the availability of antihypertensive drugs and practice guidelines, blood pressure control remains suboptimal. The complexity of current treatment guidelines may contribute to this problem. To determine whether a simplified treatment algorithm is more effective than guideline-based management, we studied 45 family practices in southwestern Ontario, Canada, using a cluster randomization trial comparing the simplified treatment algorithm with the Canadian Hypertension Education Program guidelines. The simplified treatment algorithm consisted of the following: (1) initial therapy with a low-dose angiotensin-converting enzyme inhibitor/diuretic or angiotensin receptor blocker/diuretic combination; (2) up-titration of combination therapy to the highest dose; (3) addition of a calcium channel blocker and up-titration; and (4) addition of a non—first-line antihypertensive agent. The proportion of patients treated to target blood pressure (systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg for patients without diabetes mellitus or systolic blood pressure <130 mm Hg and diastolic blood pressure <80 mm Hg for diabetic patients) at 6 months was analyzed at the practice level. The proportion of patients achieving target was significantly higher in the intervention group (64.7% versus 52.7%; absolute difference: 12.0%; 95% CI: 1.5% to 22.4%; P=0.026). Multivariate analysis of patient-level data showed that assignment to the intervention arm increased the chance of reaching the target by 20% (P=0.028), when adjusted for other covariates. In conclusion, the Simplified Treatment Intervention to Control Hypertension Study indicates that a simplified antihypertensive algorithm using initial low-dose fixed-dose combination therapy is superior to guideline-based practice for the management of hypertension.
- hypertension management
- randomized, controlled trial
- fixed-dose combination therapy
- cluster randomization
Treatment of hypertension remains suboptimal despite the development of novel therapies and the widespread implementation of education programs. Although multiple barriers exist to achieving better control of blood pressure, these mainly consist of patient- and practitioner-centered factors.
From the patient perspective, poor adherence to antihypertensive regimens is a significant component of the “treatment gap.”1–3 An important part of this problem is current prescribing practices. Unfortunately, most patients require ≥2 medications to achieve optimal control,4 and these multidrug regimens are associated with lower adherence.5,6 This has been suggested to be the most important contribution to inadequate blood pressure control.7 Furthermore, switching medications, a strategy featured prominently in most national treatment guidelines, has also been linked to poor adherence.8 Thus, the use of simpler, more effective drug regimens might improve blood pressure control.
From a practitioner perspective, a number of behavioral factors have been associated with poor blood pressure control. One of the most important factors is “therapeutic inertia,” whereby practitioners fail to appropriately escalate the intensity of therapy despite the presence of poorly controlled hypertension.9 We speculate that the increasingly complex treatment regimens currently advocated by experts, national guidelines, and the pharmaceutical industry might contribute to this undesirable behavior.
Given these issues, a simple, step-care–based algorithm for the pharmacological management of hypertension (Simplified Treatment Intervention to Control Hypertension [STITCH]) was developed. This algorithm features the initial use of a low dose of diuretic/angiotensin-converting enzyme inhibitor or diuretic/angiotensin receptor blocker fixed-dose combination. The rationale for this approach is that low-dose combinations of antihypertensive drugs are more effective than low-dose monotherapy and that the incidence of adverse effects with low-dose combination therapy is similar to that of placebo.10,11 Moreover, half-standard doses of diuretic/angiotensin converting enzyme inhibitor, diuretic/angiotensin receptor blocker combinations are less likely to disturb glucose or potassium homeostasis.10 From the patient perspective, use of fixed-dose combinations may improve adherence.12 From the physician perspective, this approach reduces the need for switching drugs, and treatment can be easily escalated. These properties may reduce therapeutic inertia and increase blood pressure control rates. However, the use of polytherapy with regard to antihypertensive efficiency has been questioned.13
Based on these considerations, we conducted a cluster-randomized, controlled trial that compared a simplified algorithm for the treatment of hypertension with conventional guideline-based care.
Study Practices and Randomization Procedure
Our hypothesis was that use of a simplified algorithm for the treatment of hypertension would result in better practice-level blood pressure control than conventional guideline-based care. The goal was to determine the effectiveness of the treatment algorithm within the family practice setting. Because this intervention was directed at the family practitioner and it would not be implementable at an individual patient level, cluster design was necessarily adopted. In addition, this cluster approach minimizes the risk of treatment contamination at the patient level. Therefore practitioners, rather than patients, were randomly assigned in a 1:1 ratio to the STITCH-care treatment algorithm or to continue with usual management according to Canadian Hypertension Education Program (CHEP) guidelines. The randomization schedule was computer generated and stratified by the year of graduation of the family physician (before 1984 or 1984 and later). Because of the nature of the intervention, participants were aware of their treatment assignment.
The study was conducted at 45 (single-physician) family practices in southwestern Ontario, Canada, between February 2005 and January 2007. Initially 950 practices were approached, and 93 physicians who expressed an interest in participation were randomly assigned to either STITCH care or to care based on the CHEP recommendations (guideline care). To avoid contamination, separate protocols were sent to practitioners assigned to STITCH care and guideline care. From these practices, eligibility to continue was based on practitioner ability to identify ≥50 patients with uncontrolled hypertension through chart review. Thirty five of the original 93 practices did not identify a sufficient number of patients and were not further evaluated. The 58 practices able to identify 50 potential subjects (and willing to continue) received their assigned intervention. Those practices that were able to identify and enroll ≥30 patients with confirmed, uncontrolled hypertension were analyzed. Because the study was analyzed at the cluster level, this was predetermined as the minimum number of subjects required to give a precise estimate of practice-level hypertension control. Participant flow is depicted in Figure 1.
At each practice, data were collected from all of the eligible patients who were willing to participate and give written informed consent to a maximum of 50 patients. Patients were enrolled consecutively as they arrived to their physician’s practice. Eligible patients were men and women with uncontrolled hypertension who were ≥18 years of age. Both untreated and partially treated patients were eligible. Uncontrolled hypertension was defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg for patients without diabetes mellitus or systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥80 mm Hg for patients with diabetes mellitus. Patients with ischemic heart disease, atrial fibrillation, peripheral vascular disease, stroke, stage 4 or 5 chronic kidney disease (ie, those with an estimated glomerular filtration rate of <30 mL/min per 1.73 m2) or those participating in other hypertension studies were not eligible.
For those partially treated patients enrolled in STITCH-care practices, physicians were not encouraged to discontinue the patient’s baseline medications but were encouraged to make subsequent drug additions using the STITCH algorithm, focusing on the use of fixed-dose combinations. Patients were not excluded for reported intolerance to any individual drug recommended in the STITCH algorithm.
The STITCH algorithm consisted of 4 steps: (1) initial therapy with a half tablet of a low-dose angiotensin-converting enzyme inhibitor/diuretic or angiotensin receptor blocker/diuretic combination; (2) up-titration of combination therapy successively to the highest dose supplied; (3) addition of a calcium channel blocker and up-titration; and (4) addition of one of the non–first-line antihypertensive agents. Practitioners assigned to the STITCH-care group were educated on the use of the treatment algorithm by an implementation team consisting of the principal investigator and study personnel from the coordinating center. Simplified aids to facilitate implementation were supplied (including treatment algorithm cards and a listing of applicable low-dose, fixed-dose combinations (Figure 2 and Table 1). The principal investigator (R.D.F.) followed up with each physician by telephone on 1 occasion, 4 to 6 weeks after the initiation visit. Practitioners assigned to the guideline care group were not exposed to the STITCH algorithm. At these practices, the CHEP guidelines (that were current at the time of the study) were reviewed with the physician, and they were provided educational materials for the management of hypertension.
The CHEP guidelines are similar in breadth and content to the Seventh Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure recommendations.14 However, in the CHEP guidelines current at the time of the study, fixed-dose combination drugs were not recommended as first-line treatment even for those with stage 2 hypertension (as they are in the Seventh Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure and in the 2008 CHEP recommendations).
For both treatment arms, the principal investigator was available to the family practitioners for consultation. Our intent was to provide equal access to expert advice and educational material to both treatment groups.
The primary outcome measure was the proportion of patients achieving target blood pressure at the practice level, as determined at a 6-month follow-up assessment. A successful treatment was defined as achieving a systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg for patients without diabetes mellitus or systolic blood pressure <130 mm Hg and diastolic blood pressure <80 mm Hg for patients with diabetes mellitus. Secondary efficacy outcomes were the change in systolic blood pressure and diastolic blood pressure. Blood pressure was measured based on the average of 5 readings, determined using a standardized and precalibrated automated blood pressure device (BpTRU, VSM Med Tech). Practices were educated in the use of the machine before initiation of the study, and detailed instructions for taking the average blood pressure measurement were attached to each machine to ensure consistent measurement across practices. All of the antihypertensive medications received by the patient during the study period were recorded on drug logs.
Assessment of total daily antihypertensive doses at the 6 month visit was based on review of drug logs. Standard dose units for hypertension treatment were assessed using our previously published approach and were generally defined as twice the recommended starting dose.15 For example, a 12.5-mg dose of hydrochlorothiazide was assessed as 0.5 standard units and a 25-mg dose assessed as 1.0 standard unit.
Practitioner Satisfaction Survey
On the completion of the study, participating physicians were sent a questionnaire assessing their satisfaction with their approach to the management of hypertension. The questionnaire was composed of the following 4 questions: (1) how effective is your treatment approach in managing your patients with hypertension?; (2) how feasible do you think it is to sustain the treatment approach in your practice setting?; (3) how likely would you be to recommend your treatment approach to a colleague?; and (4) overall, how satisfied are you with your approach to hypertension management? The questions were scored on a Likert scale, where 0 indicated the most negative response and 10 indicated the most positive response.
Descriptive statistics were used to assess the demographic characteristics of the patients and physicians. Because treatment strategies to improve the management of hypertension were perceived and implemented at the level of the practice, statistical inferences for the primary outcome were based on the proportion of patients achieving target blood pressure at the level of the practice rather than at the patient level. The primary analysis compared the proportion of patients achieving target blood pressure between the 2 treatment groups using a 2-sample t test.16 A similar approach was used to compare the other secondary outcomes. The primary outcome was also analyzed using a weighted t test.17
We also evaluated patient-level data. Factors associated with successful treatment to target blood pressure were evaluated using a modified Poisson regression approach that yields risk ratios adjusted for clustering.18,19 Provider satisfaction data were analyzed using a 2-sample t test. All of the analyses were based on the intention-to-treat principle. Statistical tests were 2 sided and were performed at the 0.05 level of significance.
Based on data obtained from a preliminary study of the effects of home blood pressure monitoring on blood pressure control,15 the interpractice SD of the proportion of patients achieving target blood pressure was estimated to be 17%. Therefore, randomization of 23 practices per group provided 80% power to detect a difference of 15% in the proportion of patients achieving target blood pressure between the treatment groups.
Ethics Approval and Informed Consent
The institutional review board at the University of Western Ontario approved the protocol. All of the patients gave written informed consent.
A total of 45 practices were successful in identifying ≥50 patients with uncontrolled hypertension and enrolling ≥30 patients with confirmed uncontrolled hypertension. Of these, 18 practices had been assigned to STITCH care and 27 had been assigned to guideline care. The physician and patient demographics were similar in the 2 treatment groups (Table 2).
A total of 2111 patients from these 45 practices were evaluated. One patient recruited to STITCH care and 6 patients from the guideline-care arm were ineligible because their baseline blood pressure was already at target. The analyses were, thus, based on data collected from 2104 patients with uncontrolled hypertension, of whom 49% were already being treated at the time of their baseline assessment. Through the duration of the study, 23 patients in the STITCH-care arm and 33 patients in the guideline-care arm withdrew from the study.
Proportion of Patients Achieving Target Blood Pressure
Patients treated at practices that were randomized to STITCH care were more likely to reach target blood pressure (Table 3⇓). At practices assigned to STITCH care, the mean success rate of achieving target blood pressure at 6 months was 64.7% compared with 52.7% for those assigned to guideline care (between-group difference: 12.1%; 95% CI: 1.5% to 22.4%; P=0.026). An analysis that weighted the t statistic for the number of patients in a practice also showed significant and similar results to the unweighted analysis.
Changes in Systolic and Diastolic Blood Pressures
After 6 months of treatment, practices assigned to STITCH care had a mean (± SD) reduction of systolic blood pressure of 22.6±4.9 mm Hg compared with 17.5±5.2 mm Hg for practices assigned to guideline care (absolute difference: 5.2 mm Hg; 95% CI: 2.0 to 8.5 mm Hg; P=0.002; Table 3⇑). The STITCH-care practices also demonstrated a greater reduction in mean diastolic blood pressure than those assigned to guideline care (10.4±3.3 versus 8.2±3.1 mm Hg; absolute difference: 2.2 mm Hg; 95% CI: 0.2 to 4.2 mm Hg; P=0.03). Seven serious adverse events were reported (3 in guideline care and 4 in STITCH care), including 3 deaths (all in the guideline-care group) and 3 episodes of drug intolerance (all in the STITCH-care group).
Exposure to Antihypertensive Drugs
Drug logs were used to assess adherence with the treatment algorithms. In the guideline-care group, 15% of patients were exposed to fixed-dose combination drugs. In contrast, in the STITCH-care group, exposure to fixed-dose combination drugs was significantly higher (85% versus 15%; P<0.001). Paralleling these findings, the number of individual drugs prescribed (with a fixed-dose combination counting as 2 individual drugs) was significantly higher in the STITCH-care practices compared with those assigned to guideline care (2.3 versus 1.9 drugs; absolute difference: 0.4 drugs; 95% CI: 0.2 to 0.6; P<0.001). These findings are consistent with the greater use of fixed-dose combinations in patients assigned to the STITCH-care arm (see Table 4).
Although more aggressive dosing might have been facilitated by the STITCH algorithm, the number of standard doses prescribed was similar in the 2 groups. In fact, the number of standard doses of antihypertensive drug prescribed tended to be higher in the guideline care arm: (1.7 versus 2.0 standard doses; absolute difference: −0.3; 95% CI: −0.5 to 0.01; P=0.06). However, it should be noted that the proportion of patients who were up-titrated during the study was higher in the STITCH-care group (82.6% versus 69.6%; between-group difference: 13.0%; 95% CI: 3.8% to 22.3%; P=0.007).
Predictors of Achieving Blood Pressure Targets
As depicted in Table 5, patient gender, patient age group, and physician graduation before 1984 were not predictors of blood pressure control at 6 months. The multivariate model identified 2 factors that were independently associated with successful treatment to target blood pressure. Assignment to the STITCH-care arm increased the probability of achieving this goal by 20% (risk ratio: 1.20; 95% CI: 1.02 to 1.40; P=0.03). Patients without diabetes mellitus had more than double the probability of reaching their target blood pressure independent of the treatment assignment. This is consistent with previous evidence of lower rates of blood pressure control in patients with diabetes mellitus (primarily ascribed to the lower targets for this subgroup of hypertensive patients).20
The physicians in the STITCH-care arm were somewhat more satisfied with their approach to managing hypertension than physicians in the guideline-care arm (8.9 versus 8.2; absolute difference: 0.7; 95% CI: 0.4 to 1.4; P=0.04). The physicians in the STITCH-care arm also tended to feel that they were more effective in managing patients with hypertension (8.70 versus 8.00; absolute difference: 0.70; 95% CI: 0.02 to 1.40; P=0.06) and were more willing to recommend their approach to other physicians (8.70 versus 7.70; absolute difference: 1.00; 95% CI: 0.02 to 2.10; P=0.05). There was no difference between the treatment group arms in their perspective of the feasibility of sustaining their current treatment approach (8.20 versus 7.90; absolute difference: 0.30; 95% CI: 0.20 to 1.10; P=0.14).
Identification of new strategies for improving blood pressure control in a family practice setting is a healthcare priority. Our study demonstrates that a simplified algorithm for the management of hypertension is implementable, changes physician-prescribing patterns, and results in better blood pressure control than conventional guideline-based care.
Notwithstanding the competing influences of both pharmaceutical company marketing and the educational initiatives of the CHEP, our simple implementation process, consisting of an initial meeting and single follow-up telephone call by a clinical pharmacologist, changed physician behavior. This behavioral change is probably best demonstrated by the ≈6-fold increase in the use of fixed-dose antihypertensive combination drugs by STITCH-care physicians in comparison with the guideline-based practices. However, most importantly, blood pressure was improved in practices assigned to the STITCH-care arm. Furthermore, STITCH-care physicians reported greater satisfaction with their management of hypertension, suggesting that STITCH care is a practicable strategy for use by community physicians.
The STITCH algorithm featured the initial use of combination therapy. It is notable that several previous studies have suggested that initial therapy with specific fixed-dose combinations results in superior blood pressure reduction in comparison with monotherapy-based approaches. However, these studies used mandated regimens in rigidly conducted clinical trials and relatively high initial drug doses.21,22 In contrast, our trial used a “critical pathway” approach to educating physicians regarding a simplified treatment algorithm that was not dependent on the prescription of specific drug combinations. Accordingly, we believe that these findings are more widely generalizable for the management of hypertension.
Which components of STITCH were responsible for the effectiveness of the algorithm? Although this question was not directly addressed in the study, 3 factors should be considered. First, use of the algorithm might have facilitated the use of more intensive dosing regimens by physicians (ie, overcoming therapeutic inertia). Second, use of fixed-dose combinations may have improved adherence by patients, as has been suggested in observation studies. Finally, use of combination therapy may be intrinsically more effective because of synergy between agents. However, with respect to the first possibility, our analysis of dose intensity showed that practices assigned to STITCH care did not receive more aggressive treatment, as assessed by the number of standard doses prescribed. On the other hand, patients in the STITCH-care arm were more likely to have had their antihypertensive dose up-titrated. Therefore, whether the STITCH algorithm attenuated therapeutic inertia cannot be resolved. Thus, we speculate that the effectiveness of STITCH is more likely related to better adherence by patients and/or the use of more effective drug combinations (ie, minimizing the use of less-than-additive 2- or 3-drug combinations, a phenomenon suggested to occur with polytherapy in hypertension).13
The 2008 CHEP guidelines included a new recommendation supporting the use of initial therapy with a fixed-dose combination as a first-line choice. Notably, based on the 2008 CHEP recommendations, 45% of STITCH-care and 39% of guideline-care patients would have been candidates for initial fixed-dose combination therapy at baseline. However, it is important to emphasize that starting with a fixed-dose combination is only 1 of 6 first-line recommended choices for initiation of therapy in this group. We would suggest that, based on our findings, a fixed-dose combination might be the preferred initial treatment in this subgroup (versus initiation with a single agent).
Our study has some limitations. The STITCH algorithm was implemented by a single clinical pharmacologist (R.D.F.). Whether similar results could be obtained with widespread implementation is unknown. Second, the imbalance between numbers of practices initiated in STITCH-care versus guideline-care arms requires consideration. This imbalance raises the possibility that the physicians who participated in the STITCH-care arm may have been different, in terms of their behavior, from those assigned to the GUIDELINE-care arm. However, the majority of the practice “dropouts” (35 of 48 practices) occurred before site initiation (ie, before the point at which intervention sites were introduced to the treatment algorithm). Among practices that actively started screening, dropout rates were comparable. Second, we could not identify any practice or patient level factor that would support the existence of a difference between treatment arms. Third, the treatment effect size was relatively large. We do not believe it is plausible that confounding would result in a difference of this magnitude. Moreover, the trial was designed with a 1:1 ratio between STITCH care and guideline care, which, in general, has higher power than an unbalanced trial. The unbalanced ratio of intervention:control sites in STITCH would have been expected to reduce the power of the study to identify significant differences. This implies that our results are likely conservative.
Use of a simplified algorithm for the treatment of hypertension featuring the initial use of low-dose, fixed-drug combination drugs is implementable, accepted by family physicians, and results in improved blood pressure control. This return to step care may be an important way forward for the treatment of hypertension. We also speculate that such an approach may be broadly applicable for the treatment of other chronic diseases.
Coordinating center personnel include the following individuals: M. Abboud, B. Annunziello, T. Clayton, B. Feagan, R. Feldman, L. Jensen, E. Karugu, H. Kaufman, C. Matthews, J. MacKay, K. Greenwood, S. Nelson, N. Ng, W. Piper, L. Robinson, B. Sarazin, C. Trinidad, M. Vandervoort, C. Wong, J. Xie, and G. Y. Zou.
The following family practitioners participated in the study: Ancaster: B. B. Loewith; Brantford: W. S. Tsuchida, D. C. Vincent; Burlington: R. R. Jhirad, A. Nayar; Chatham: J. Boekhoud, S. Munro; Exeter: D. Hodder; Hamilton: R. D. Arora, T. Subramanian, L. S. Zavodni, Ilderton: G. Nancekievill; Kitchener: K. Bedrosian, B. Huth, J. A. King, S. MacGregor, R. Nicholson, T. Szozda; London: T. Alam, E. Arcia-Bravo, E. Armogan, V. Chawla, G. Debosz-Szczerbowski, P. Dickie, P. Dzongowski, J. Horne, S. Joshi, R. Komar, A. Mawji, S. Petrlich, A. Richard, F. Wong; St Mary’s: S. Hiscock; St Thomas: B. Hertwig; Stratford: D. Crowley; Strathroy: M. Mithoowani; Thedford: L. Gibbs; Waterloo: F. Dawood, G. Pannozzo, J. Peet; Windsor: M. Bilibajkich, D. Bridgeo, M. Zajner; and Woodstock: J. Brioux.
We acknowledge the contribution of Suzanne Vandervoort, the study’s initial coordinator. We thank Beverley Jasevicius for expert secretarial assistance.
Sources of Funding
The trial was supported by an independent grant from Pfizer Canada.
The sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation review; or approval of the article.
R.D.F. reports receiving research funding from Pfizer and has been a paid consultant or on advisory boards for Bayer, Boehringer-Ingleheim, Pfizer, Bristol Myers Squibb, Sanofi-Aventis, Schering-Plough, Merck, and Servier. B.G.F. reports receiving research funding unrelated to the topic of the article from Synta, Millennium, Schering Canada, Celltech, Centocor, Elan/Biogen, Berlex, Ortho-Biotech, Protein Design Labs, ISIS, Santarus, Schering Plough, Celgene, UCB Pharma, Napo Pharma, BMS, Abbott, and Otsuka. He also reports receiving consulting and lecture fees from UCB Pharma, Schering Canada, Proctor and Gamble, Elan/Biogen, Millennium, Protein Design Labs, Berlex, AstraZeneca, Celgene, Abbott, Santarus, GeneLogic Inc, Cerimon Pharmaceuticals, Tioga Pharmaceuticals, BMS, ISIS, Serono, Teva, Genentech, and CombinatoRx. No other financial disclosures were reported.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
This trial has been registered at www.clinicaltrials.gov (identifier NCT00129909).
- Received September 22, 2008.
- Revision received October 8, 2008.
- Accepted January 28, 2009.
Cushman WC, Ford CE, Cutler JA, Margolis KL, Davis BR, Grimm RH, Black HR, Hamilton BP, Holland J, Nwachuku C, Papademetriou V, Probstfield J, Wright JT Jr, Alderman MH, Weiss RJ, Piller L, Bettencourt, J, Walsh SM; ALLHAT Collaborative Research Group. Success and predictors of blood pressure control in diverse North Am settings: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). J Clin Hypertens. 2002; 4: 393–404.
Caro JJ, Speckman JL, Salas M, Raggio G, Jackson JD. Effect of initial drug choice on persistence with antihypertensive therapy: the importance of actual practice data. CMAJ. 1999; 160: 41–46.
Okonofua EC, Simpson KN, Jesri A, Rehman SU, Durkalski VL, Egan BM. Therapeutic inertia is an impediment to achieving the Healthy People 2010 blood pressure control goals. Hypertension. 2006; 47: 345–351.
Law MR, Wald NJ, Morris JK, Jordan RE. Value of low dose combination treatment with blood pressure lowering drugs: analysis of 354 randomised trials. BMJ. 2003; 326: 1427.
Khan NA, Hemmelgarn B, Padwal R, Larochelle P, Mahon JL, Lewanczuk RZ McAlister FA, Rabkin SW, Hill MD, Feldman RD, Schiffrin EL, Campbell NR, Logan AG, Arnold M, Moe G, Campbell TS, Milot A, Stone JA, Jones C, Leiter LA, Ogilvie RI, Herman RJ, Hamet P, Fodor G, Carruthers G, Culleton B, Burns KD, Ruzicka M, deChamplain J, Pylypchuck G, Gledhill N, Petrella R, Bouldanger JM, Trudeau L, Hegele RA, Woo V, McFarlane P, Touzy RM, Tobe SW; Canadian Hypertension Education Program. The 2007 Canadian Hypertension Education Program recommendations for the management of hypertension: part 2 - therapy. Can J Cardiol. 2007; 23: 539–550.
Zarnke KB, Feagan BG, Mahon JL, Feldman RD. A randomized study comparing a patient-directed hypertension management strategy with usual office-based care. Am J Hypertens. 1997; 10: 58–67.
Bland JM, Kerry SM. Statistical notes: weighted comparisons of means. BMJ. 1998; 316: 129.
Zou GY. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004; 159: 702–706.
Zou GY. One relative risk versus two odds ratios: implications for meta-analyses involving paired and unpaired binary data. Clin Trials. 2007; 4: 25–31.
Mourad JJ, Waeber B, Zannad F, Laville M, Duru G, Andréjak M, Investigators of the STRATHE Trial. Comparison of different therapeutic strategies in hypertension: a low-dose combination of perindopril-indapamide versus a sequential monotherapy or a stepped-care approach. J Hypertens. 2004; 22: 2379–2386.