Abstract
Keywords
Type 2 diabetes is increasing in prevalence worldwide and is a major contributor to morbidity and mortality globally. 1 Patients with type 2 diabetes are at increased risk for microvascular complications, with type 2 diabetes being the leading cause of end-stage renal disease. 2 Moreover, morbidity and mortality due to macrovascular complications are 2 to 5 times higher in people with diabetes,3,4 which imparts a major economic burden on patients and the health system. 5 Intensified, multifactorial interventions using behavioral and pharmacological therapy targeting modifiable risk factors such as hypertension, dyslipidemia, and hyperglycemia may reduce both microvascular and macrovascular complications by up to 50%.6-8
Previously, we have demonstrated that a similar intervention promoting healthy behaviors and applying pharmacological algorithms can be delivered by local multidisciplinary teams in a community setting to improve blood pressure, glycemic control, and dyslipidemia, utilizing nonphysician staff to maintain regular follow-ups and ensure adherence with pharmacotherapy and lifestyle changes. 9 In general, such interventions are evaluated by examining whether targets are achieved for individual outcomes for the study population as a whole. While this can be useful to determine whether an intervention is successful, it may overlook subgroups of individuals for whom the intervention is less effective or aspects of the intervention that need to be revised. To refine and improve delivery of effective care in clinical practice, an evaluation focusing on the effectiveness of an intervention for individual patients would be helpful.
Given the benefits of multiple risk factor–targeted interventions for high-risk individuals with type 2 diabetes, it is important to optimize the delivery of proven interventions in a real-world setting. Therefore, we examined whether type 2 diabetic patients enrolled in community-based clinics uniformly benefited from multifactorial interventions and examined the characteristics of patients who did not achieve targets, with the objective of identifying potential barriers to success.
Materials and Methods
We conducted a retrospective review in high-risk type 2 diabetic patients with hypertension and/or albuminuria attending 5 diabetic nephropathy prevention clinics (DNPCs) in Northern Alberta, Canada. The rationale, design, and structure of the clinics have been described previously. 9 Each clinic is staffed by a nurse and a dietitian and supported by a central team comprising a pharmacist, nephrologist, and endocrinologist. The goals of the clinic are to promote lifestyle change (smoking cessation, increased physical activity, healthy diet, weight loss) and promote the use of proven pharmacotherapies (angiotensin-converting enzyme [ACE] inhibitors, statins, aspirin and help reach targets for blood pressure, lipids, and A1c) by previously published algorithms. 9
This exploratory report focuses on type 2 diabetic patients attending between January 1, 2004 and August 31, 2007 who had completed a minimum of 1 year of follow-up and who had a complete data set. For each individual, the attainment of 7 risk factor targets was evaluated: (1) blood pressure (<130/80 mm Hg), (2) A1c (<7% or 6.5% if lifestyle managed), (3) low-density lipoprotein (LDL) cholesterol (<2.5 mmol/L [<96 mg/dL]), (4) weight loss (>10% of original weight), (5) exercise (30 minutes, 3 times per week), (6) a diabetic meal plan (dietitian advised), and (7) smoking status (nonsmoker). Ethics approval was obtained from the University of Alberta Health Research Ethics Board.
Statistical Analysis
The overall effectiveness of the DNPCs was assessed by comparing the percentage of patients who achieved clinical risk factor targets at follow-up compared to baseline. Patients were divided into groups based on the number of targets achieved: group 1 or less successful (0-2 targets), group 2 or moderately successful (3-4 targets), and group 3 or successful (≥5 targets). Target cut points for the groups were determined prior to analysis based on the frequency distribution of the number of targets achieved with an effort to maximize the number of patients within 3 groups. Moreover, no attempt was made to weigh the clinical importance of each target, and all targets were considered equally important. We acknowledge these cut points are arbitrary but likely represent a real-world metric used by clinicians to measure success in the management of patients with type 2 diabetes. Moreover, similar cut points have been used in previous studies of multifactorial risk factor management programs. 8 The data presented represent a post hoc exploratory analysis, and no a priori power calculation had been performed. The 3 groups were compared using χ2 tests, Wilcoxon signed-rank test, or analysis of variance as appropriate. All statistical analysis was conducted using Stata/SE version 11.0 (StataCorp LP, College Station, Texas).
Results
Overall, 522 patients had been followed regularly for at least 1 year between January 1, 2004 and August 31, 2007. Of these, 235 patients had complete data both at baseline and 1 year to permit analysis of all 7 risk factor targets. Patients were 62 (range, 28-87) years of age and had diabetes of 9 years’ (SD, 9) duration. There was a slight male preponderance (57%), and 181 (77%) reported a history of cardiovascular disease at the initial visit. The baseline characteristics (age, sex, duration of diabetes, renal function, or cardiovascular history) did not differ from patients with incomplete data who were not included in this report (P > .05 for all comparisons). The clinical characteristics at baseline and follow-up are presented in Table 1.
Characteristics of 235 Patients With Type 2 Diabetes Followed for 1 Year at Community-Based Diabetes and Nephropathy Clinics in Alberta, Canada
Note: Data are mean (standard deviation [SD]) or n (%). GFR, GFR, glomerular filtration rate; LDL, low-density lipoprotein; HDL, high-density lipoprotein; ACE, angiotensin-converting enzyme; ARB, ARB, angiotensin receptor blocker; CCB, CCB, calcium channel blocker.
There was a significant increase in the number of patients reaching targets for blood pressure, HbA1c, and LDL (P < .001 for all) (Table 1 and Figure 1). There were significant reductions in mean blood pressure (4/3 mm Hg), HbA1c (0.4%), and LDL (0.3 mmol/L) (P < .001 for all) (Table 1). There was no change in weight or serum creatinine (Table 1). The proportion of patients making positive lifestyle changes also increased, with more patients reporting adhering to a dietitian-advised meal plan (P = .006), exercise program (P = .001), and quitting smoking (P < .001) (Table 1 and Figure 1).
Proportion of patients meeting risk factor targets at baseline and 1-year follow-up at community-based diabetic nephropathy prevention clinics in Alberta, Canada.
Of 6 possible risk factor targets measured at baseline (weight loss target is not measurable at baseline), 100 (43%) patients had met 0 to 2 targets, 123 (52%) patients had met 3 to 4 targets, and 12 (5%) patients had met 5 risk factor targets. After 1 year, the proportion achieving little success (0-2 targets) had decreased (n = 47, 20%), while the proportions achieving higher degrees of success increased (123 [52%] patients achieved 3-4 targets, and 12 [5%] achieved ≥5 risk factor targets) (Figure 2).
Distribution of patients by number of targets achieved at baseline and after 1 year of follow-up at community-based diabetic nephropathy prevention clinics in Alberta, Canada.
The clinical characteristics of these 3 groups (defined by the number of targets reached) are presented in Table 2. Although at baseline age, weight, blood pressure, use of antihypertensives, and dietary and exercise habits were similar between the groups, there were important differences in gender, A1c, LDL, smoking habits, and diabetes management. The group of successful patients contained more females (P ANOVA = .003), had shorter diabetes duration (P ANOVA = .03), and tended to be less likely to use insulin (P ANOVA = .06). The group of patients who were “unsuccessful” had a higher proportion of smokers (P ANOVA < .001) and, as might be expected, had higher A1c (P ANOVA = .02) and LDL (P ANOVA = .007).
Demographic, Clinical, and Biochemical Characteristics of 235 Patients Achieving 0 to 2, 3 or 4, and 5 or Greater Clinical Targets
Note: P values represent ANOVA for continuous variables and χ2 for categorical variables with 2 degrees of freedom. SD, standard deviation; BP, blood pressure; LDL, low-density lipoprotein.
There were striking differences between the groups in their responses to the intervention. Rather than improving, patients in group 1 (“unsuccessful”), on average, had an increase in systolic blood pressure (2.1 ± 17.6 mm Hg), while patients in groups 2 and 3 had a reduction in systolic blood pressure (-4.4 ± 16.4 mm Hg and −7.2 ± 14.2 mm Hg, respectively), despite similar numbers of antihypertensive medications being used (Table 2 and Figure 3A). Diastolic blood pressure tended to decrease in all 3 groups, although numerically greater reductions were seen in the “successful” group (Table 2 and Figure 3A). Although there were statistically significant differences in A1c between the groups at follow-up, the reduction in A1c was similar in all 3 groups (Table 2 and Figure 3B). Diabetes management, however, was more intensive among patients in group 1 (“unsuccessful”) as they were treated more aggressively both at baseline and follow-up (Table 2). For example, at follow-up, 21 (45%) patients in group 1 were using insulin compared to 45 (34%) patients in group 2 and 10 (18%) patients in group 3 (P ANOVA = .03) (Figure 4). The LDL cholesterol levels at follow-up were lower in all 3 groups, but the reduction was greatest in group 3 (Figure 3C). Group 3 was the only group in which patients lost weight, and these patients were more likely to report adhering to an exercise regimen or a meal plan at follow-up. There were no smokers in group 3 at follow-up. Six smokers quit between baseline and follow-up in groups 2 and 3, while 1 ex-smoker relapsed in group 1.
Changes in blood pressure, HbA1c, and low-density lipoprotein (LDL) in all 3 groups.
Proportion of patients in each group managing their diabetes with diet, oral antihyperglycemic agents (OHAs), or insulin ± OHAs at follow-up.
By definition, the proportion of patients reaching targets was low in group 1, but the targets reached most frequently were report of adherence to a meal plan (55%) and nonsmoking (53%), followed by LDL cholesterol (32%), blood pressure (19%), and A1c (11%), while only 2 patients (4%) reported participating in exercise, and no patients achieved weight loss of 10%. In group 2, although the proportions were higher, the ranking of likelihood to reach targets followed a very similar sequence: nonsmokers (92%), followed a meal plan (82%), LDL (70%), blood pressure (50%), A1c (41%), exercise plan (21%), and weight loss (2%). A similar sequence was seen in group 3: nonsmokers (100%), meal plan (95%), LDL (88%), blood pressure (86%), A1c (70%), exercise regimen (70%), and weight loss (20%).
Discussion
Multifactorial risk factor reduction programs decrease mortality as well as microvascular and macrovascular complications in high-risk diabetic individuals,6,7 although their application in a real-world setting is less certain. In our community-based programs, following 1 year of participation, the number of patients achieving predefined risk factor targets increased significantly. However, one fifth of patients had limited success, achieving targets for ≤2 of the 7 parameters. These unsuccessful patients were more likely to be male, smokers, had a longer duration of diabetes, worse glycemic control, and a higher initial LDL.
It might be assumed that the lack of success simply reflected differences in baseline parameters, so patients with higher initial LDL would be less likely to reach the target. Although this may be true to some extent, there were striking differences between successful and unsuccessful patients in their response to the intervention, as indicated by the changes between baseline and follow-up. While the reduction in A1c (0.5%) was similar in all groups, changes in systolic blood pressure, LDL, and weight differed substantially, with clear improvements in the “successful” group but little change or worsening observed in the “unsuccessful” group. This was despite similar use of pharmacotherapy for blood pressure and greater use of insulin among unsuccessful patients. Interestingly, the proportion of unsuccessful patients reporting positive lifestyle changes (not smoking, following a meal plan, participating in exercise) was much lower compared with intermediate or successful patients. These observations raise the intriguing suggestion that the intervention was not equally effective for all patients and that effectiveness of pharmacotherapy may be heavily influenced by lifestyle factors.
Our data are similar to previous studies,6,7 which have found that it is easiest to achieve targets for LDL, followed by blood pressure, but that achieving glycemic targets is most difficult. This hierarchy was similar across the 3 groups in our study. In our study, the improvement in A1c did not differ between groups, suggesting that the effectiveness of this aspect of our intervention was equally effective across groups.
There are a number of potential explanations for the limited success of some patients. Longer duration, more severe diabetes might make achievement of targets difficult. Moreover, the higher proportion of male patients in the “unsuccessful” group may reflect gender differences in the ability to implement lifestyle changes or in the effectiveness of our predominantly female clinic staff to deliver this message to male participants. Indeed, this group was less likely to exercise or adhere to a meal plan and tended to gain weight. This suggests that poor adherence to clinic recommendations might be an important explanation because resistant hypertension would be expected if recommendations to restrict dietary sodium were ignored. Alternatively, the greater utilization of insulin in this group may have predisposed to weight gain and secondarily led to resistant hypertension. 10
The effectiveness of multifactorial interventions is often used to justify the promotion of “proven” pharmacotherapies with less emphasis on the lifestyle components of these interventions. Our findings are similar to those of a Swedish study documenting that, with extended follow-up, biochemical parameters improved; however, adverse lifestyle characteristics such as increased waist circumference and lack of exercise persisted. 11 This suggests that regardless of intense pharmaceutical intervention, lifestyle nonadherence may continue to offset some of the gains made by increased pharmacotherapy. Lifestyle interventions, however, tend to be expensive and difficult to replicate outside the setting of clinical trials. 12 Our findings suggest that an approach effectively combining healthy lifestyle changes with pharmacotherapy may achieve optimal results.
Although this study represents a novel approach to exploring variability in responses to this multifactorial intervention in a real-world setting, it has some weaknesses. It is encouraging that the improvements in our previous report 9 have now extended beyond 1 year, and we can only speculate whether improvements will persist long term. In addition, this analysis is restricted to those patients who returned for follow-up and may represent more motivated patients. However, the baseline characteristics of the patients included in the analysis do not differ significantly from those who did not return for follow-up. Further, a number of data elements relied on self-report, and the data were gathered in the setting of routine clinical care; thus, the quality of data is likely inferior to that gathered in a clinical trial. Nevertheless, this is representative of the real-world data that clinic personnel use to effectively guide therapy directed to manage risk factors.
Overall, we have demonstrated the general effectiveness of a community-based multidisciplinary care model that provides intense multifactorial risk reduction through the use of standardized protocols. We have identified that this approach is less successful in some patients whose failure to make lifestyle modifications is associated with resistant hypertension and inability to reach risk factor targets despite escalation of pharmacological measures. These patients may be subject to a further increase in pharmacological therapy when perhaps greater benefit would be derived from a focus on facilitating lifestyle modifications. Further study of participants in multiple risk factor management programs is required to explore additional factors affecting target achievement such as the role of ethnicity, educational level, literacy, fluency in English, socioeconomic status, and psychosocial factors. Identification of additional potential barriers may in turn help to develop strategies to further optimize care for patients with type 2 diabetes.
Footnotes
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.