Volume 118, Issue 12 , Pages 1414.e21-1414.e27, December 2005
Effects of structured care by a pharmacist-diabetes specialist team in patients with Type 2 diabetic nephropathy
Article Outline
Abstract
Purpose
We developed a disease management program for patients with Type 2 diabetic nephropathy and evaluated its effects on the time to onset of end-stage renal disease or all-cause death compared with usual care.
Methods
In a 2-year, prospective cohort study, we compared the clinical outcomes of patients managed by a structured care protocol (n = 80) to a group receiving usual care (n
=80) in the same hospital. Patients aged ≤80 years with type 2 diabetes, serum creatinine 150-400 μmol/L, and micro- or macroalbuminuria were recruited. The structured care protocol was implemented by a pharmacist-diabetes specialist team with particular emphasis on periodic laboratory assessments, patient adherence, risk factors control, and use of renin-angiotensin system inhibitor. The primary endpoint was the composite of end-stage renal disease or all-cause death. Other endpoints were the rate of renal decline, processes-of-care measures, and control of risk factors.
Results
During 22.8±7.9 months of follow-up, the primary endpoint developed in 24 and 40 patients in the structured care and usual care groups, respectively (adjusted risk reduction, 60%, P< .001). Structured care (hazard ratio [95% confidence interval (CI)], 0.40 [0.23-0.68]), age (0.95 [0.93-0.98]), baseline systolic blood pressure (BP) (1.014 [1.003-1.026]), logarithm (base 10) of baseline serum creatinine (34 441 [2290-517915]), and macroalbuminuria (8.95 [1.22-65.38]) were independent predictors for the primary endpoint. Structured care slowed the rate of renal decline (P=.032). More intensive laboratory measurements, increased use of renin-angiotensin system inhibitor, and greater reductions in BP and low-density lipoprotein (LDL) cholesterol were reported by patients receiving structured care.
Conclusions
Structured care delivered by a pharmacist-diabetes specialist team reduced the incidence of end-stage renal disease or death compared with usual care in patients with type 2 diabetic nephropathy.
Keywords: Structured care , Diabetic nephropathy , Disease management , Renoprotection , Health care delivery
Type 2 diabetic nephropathy is the leading cause of end-stage renal disease worldwide.1 In clinical trials, the progression of diabetic nephropathy to end-stage renal disease was found to be largely preventable by aggressive treatment of high blood pressure (BP),2 hyperglycemia,3 and dyslipidemia,4 as well as use of inhibitor of the renin-angiotensin system.5, 6 In routine practice, however, these strategies have not been effectively implemented.7, 8
Disease management, a systematic approach to patient care along the continuum of disease and across healthcare delivery systems, is gaining popularity in management of chronic conditions.9 No formal disease management programs for treating patients with type 2 diabetic nephropathy have been reported in the literature.10, 11 We have developed a structured care model for managing this group of patients. In this study, we examined the effects of this program on renal outcomes and processes of care.
Methods
Design
This 2-year prospective cohort study compared patients receiving structured care versus those receiving usual care. The study was conducted at the Prince of Wales Hospital, a university-based, public hospital in Hong Kong.
In June 2000, the Diabetic Nephropathy Clinic was established as a sub-clinic of the Prince of Wales Hospital diabetes clinic by a project team consisting of diabetes specialists and a pharmacist. Eighty eligible patients attending the diabetes clinic were identified from the Prince of Wales Hospital computerized database and by review of medical records. They were invited to enter the Diabetic Nephropathy Clinic to receive structured care. Because the program was service-based and delivered as a quality improvement project, approval from Institutional Review Board was not required.
The usual care arm consisted of 80 patients who, identified in the same way, met the same eligibility criteria and who were attending Prince of Wales Hospital medical clinics for management of diabetic nephropathy. They continued to receive usual care at their original clinics during the same period. They were prospectively observed as in a clinical audit and were not subject to any special interventions. The patients were not aware of the study.
Patients in both groups were followed until study end or development of primary endpoint, when data for all patients were censored.
Patients
The study involved Chinese, type 2 diabetic patients, aged 30-80 years, who had micro- or macroalbuminuria, and serum creatinine concentration 150-400 μmol/L (1.7-4.5 mg/dL), with no evidence of rapid renal progression as defined by an increase in serum creatinine by >50% in the last 6 months. Micro- and macroalbuminuria were defined according to recommendations of the American Diabetes Association.12 Patients were excluded if they had ultrasonographic evidence of obstructive uropathy, nondiabetic renal diseases, marked dementia or other unstable psychiatric illnesses, or malignancy or other life-threatening conditions.
Structured care
Details of the structured care protocol were as follows.
Usual care
Patients in the usual care group were managed according to the protocol of the clinic they were attending. The majority of doctors who saw these patients were nondiabetes medical specialists or general internists. Patients tended to see different doctors at each visit. The frequencies of clinic visits and laboratory tests as well as attainment of treatment targets were subject to individual doctors’ discretion. Pharmacists were not involved in patient care except for dispensing.
Outcome measures
The primary endpoint was the time to the first event of the composite of end-stage renal disease or all-cause death. End-stage renal disease was defined as the permanent need for dialysis or renal transplantation or serum creatinine >500 μmol/L (5.7 mg/dL), as confirmed by a second measurement taken at least 2 weeks apart. The secondary endpoint was the progression of renal disease indicated by the rate of change in serum creatinine concentration. Process-of-care measures (frequencies of laboratory measurements, and use of renin-angiotensin system inhibitor and statin) and changes in risk factors control were analyzed as tertiary endpoints.
Statistical analyses
The analyses were based on an intention-to-treat approach. All hypotheses were tested at a 2-sided significance level of 5%. The time-to-event variables, including the primary composite endpoint and its 2 components, were analyzed by Kaplan-Meier procedure.13 The effects of structured care (ie, hazard ratio and its 95% confidence interval [CI]) were estimated using the Cox proportional hazards model14 with all measurable confounders adjusted (Table 1). The number needed to treat (95% CI) was calculated from estimates of survival probabilities and hazard ratios (95% CI).15
Table 1. Covariates predefined and measured for adjustment of the effect of structured care on the primary composite endpoint and its components
| Demographics |
| Age, sex, and duration of diabetes |
| Baseline clinical characteristics |
| Presence of diabetes-related complications, body mass index, and systolic and diastolic blood pressure |
| Baseline biochemical characteristics |
| Serum creatinine concentration, HbA1c, serum triglycerides, serum LDL cholesterol, blood hemoglobin, and albuminuric status |
| Baseline drug usage |
| Use of renin-angiotensin system inhibitor, statin, and insulin |
| Level of healthcare resources utilization in the previous 2 years |
| Frequencies of hospitalizations and emergency room visits, and length of hospital stay |
For the analysis of renal disease progression, we used a linear random-effects model to compare the rate of change in serum creatinine concentration in the two care groups.16 Data were censored for patients who had reached the primary endpoint. Measurements taken during hospital admissions were excluded from analysis.
Sample size calculation
In a prior cohort of 87 patients with type 2 diabetes and serum creatinine 150-400 μmol/L attending the outpatient medical clinics of our hospital during 1996-2000, the annual incidence of death, end-stage renal disease, or doubling of serum creatinine was 37.9% (data unpublished). We estimated that the primary composite endpoint of death or end-stage renal disease would occur at a rate of 30% per year in the usual care group. For a 2-year study, and assuming a 40% risk reduction with structured care, 69 patients would be required for each group to achieve 80% power at a 5% 2-sided alpha level based on the proportional hazards model. Allowing for patients lost to follow-up, 80 patients were recruited into each group.
Results
Between June 2000 and January 2001, 80 eligible patients were invited to enter the Diabetic Nephropathy Clinic. All 80 patients agreed, although one subject refused the regular interim visit by pharmacist due to difficulty in taking leaves from work. He was also included in the structured care group in the final analyses. All 80 patients were followed at the Diabetic Nephropathy Clinic until study end or when they died or reached end-stage renal disease.
Among the 80 patients in the usual care arm, 5 were referred to other public hospitals or community-based medical clinics during the study period. We were able to determine their clinical and survival status using the hospital computerized database or by telephone contact at study end.
Baseline characteristics were similar in the 2 groups, although patients in the structured care group had higher body mass index and HbA1c and a higher prevalence of neuropathy. Patients receiving structured care had more frequent use of insulin (Table 2). The mean ± standard deviation follow-up period was 22.8±7.9 months (structured care, 24.4±7.4; usual care, 21.3±8.1 months). The study was completed in January 2002 when the last recruited patient had been followed for 2 years.
Table 2. Baseline characteristics between the structured care and usual care groups
| Overall (n = 160) | Structured care (n = 80) | Usual care (n = 80) | P value | |
|---|---|---|---|---|
| Demographics | ||||
| Age (years) | 65.1±8.8 | 64.5±9.7 | 65.8±7.8 | .354 |
| Male⁎ | 94(58.8%) | 50(62.5%) | 44(55.0%) | .335 |
| Duration of diabetes (years) | 12.3±7.2 | 13.0±6.9 | 11.4±7.5 | .205 |
| Systolic blood pressure (mm Hg) | 151.8±24.8 | 151.2±26.2 | 152.5±23.4 | .729 |
| Diastolic blood pressure (mm Hg) | 74.8±13.1 | 75.7±11.6 | 73.9±14.5 | .404 |
| Body mass index (kg/m2) | 25.4±4.1 | 26.2±4.1 | 24.7±3.9 | .021 |
| Medical history⁎ | ||||
| Retinopathy | 121(75.6%) | 63(78.8%) | 58(72.5%) | .357 |
| Neuropathy | 86(53.8%) | 53(66.3%) | 33(41.3%) | .002 |
| Peripheral vascular disease | 22(13.8%) | 15(18.8%) | 7(8.8%) | .066 |
| Lower extremity amputation | 8(5.0%) | 5(6.3%) | 3(3.8%) | .468 |
| Ischemic heart disease | 29(18.1%) | 13(16.3%) | 16(20.0%) | .538 |
| Myocardial infarction | 7(8.8%) | 2(2.5%) | 5(6.3%) | .147 |
| Coronary revascularization procedure | 12(7.5%) | 6(7.5%) | 6(7.5%) | 1.000 |
| Congestive heart failure | 21(13.1%) | 10(12.5%) | 11(13.8%) | .815 |
| Stroke | 31(19.4%) | 16(20.0%) | 15(18.8%) | .841 |
| Smoking | .834 | |||
| Current | 24(15.1%) | 11(13.8%) | 13(16.5%) | |
| Ex-smoker | 47(29.6%) | 23(28.8%) | 24(30.4%) | |
| Never | 88(55.3%) | 46(57.5%) | 42(53.2%) | |
| Alcohol | .182 | |||
| Current | 9(5.7%) | 3(3.8%) | 6(7.6%) | |
| Ex-drinker | 37(23.3%) | 23(28.8%) | 14(17.7%) | |
| Never | 113(71.1%) | 54(67.5%) | 59(74.7%) | |
| Biochemical | ||||
| Serum creatinine (μmol/L)† | 207.2(199.6to215.2) | 206.3(195.0to218.3) | 208.2(197.8to219.0) | .817 |
| Serum urea (mmol/L) | 13.8±4.5 | 13.7±4.5 | 13.9±4.4 | .823 |
| Serum cholesterol (mmol/L) | ||||
| Total | 5.6±1.4 | 5.6±1.3 | 5.6±1.6 | .969 |
| LDL-C | 3.4±1.2 | 3.3±1.1 | 3.4±1.3 | .834 |
| HDL-C | 1.2±0.4 | 1.2±0.3 | 1.2±0.4 | .854 |
| Serum triglycerides (mmol/L)† | 2.0(1.8to2.2) | 2.0(1.7to2.3) | 2.1(1.8to2.4) | .608 |
| Blood hemoglobin (g/dL) | 12.0±2.0 | 12.2±1.9 | 11.6±2.0 | .055 |
| Fasting plasma glucose (mmol/L) | 8.1±2.7 | 8.1±2.8 | 8.1±2.5 | .858 |
| Glycosylated hemoglobin (%) | 7.6±1.5 | 7.8±1.3 | 7.3±1.6 | .020 |
| Microalbuminuria⁎ | 27(16.9%) | 16(20.0%) | 11(13.8%) | .291 |
| Macroalbuminuria⁎ | 133(83.1%) | 64(80.0%) | 69(86.3%) | .291 |
| Patterns of drug usage | ||||
| Use of ACE inhibitor or AII antagonist⁎ | 103(64.4%) | 54(67.5%) | 49(61.3%) | .409 |
| Use of insulin⁎ | 89(55.6%) | 58(72.5%) | 31(38.8%) | <.001 |
| Use of statin⁎ | 51(31.9%) | 31(38.8%) | 20(25.0%) | .062 |
| Healthcare resources utilization in previous 2 years‡ | ||||
| Number of emergency room visits | 0.5(1.0) | 0.5(1.0) | 0.5(1.0) | .880 |
| Number of all-cause hospital admissions | 1.0(2.0) | 1.0(2.0) | 1.0(2.0) | .603 |
| Length of hospital stay (days) | 5.0(14.0) | 6.5(12.8) | 4.0(15.0) | .975 |
⁎ Number of patients (%). |
† Geometric mean (95% confidence interval). |
‡ Median (IQR). |
Primary composite endpoint
There were 24 and 40 events of the primary composite endpoint in the structured care and usual care groups, respectively (incidences, 14.8 vs 28.2 per 100 patient-years). The rates of end-stage renal disease were 13.5 (n=22) and 24.0 (n=34) per 100 patient-years in the two groups, respectively. The corresponding all-cause death rates were 4.3 (n=7) and 14.8 per 100 patient-years (n=21). The causes of death were end-stage renal disease (n=4), cardiovascular (n=1), and others (n=2) for the structured care group; and end-stage renal disease (n=11), cardiovascular (n=8), and others (n=2) for the usual care group.
Figure 1 displays the Kaplan-Meier plots for the primary composite endpoint (panel A) and its 2 components (panels B and C). The Cox proportional hazard model found that structured care (hazard ratio [95% CI], 0.40 [0.23-0.68], P <.001), age (0.95 [0.93-0.98], P <.001), baseline systolic BP (1.014 [1.003-1.026], P=.015), logarithm (base 10) of baseline serum creatinine (34441 [2290-517915], P <.001), and macroalbuminuria (8.95 [1.22-65.38], P=.031) were independent predictors for the primary composite endpoint. Patients in the structured care group had an adjusted relative risk (95% CI) of reaching the primary endpoint that was 60% (32-77%) lower than that in the usual care group. The relative risk reductions (95% CI) for end-stage renal disease and all-cause death were 55% (20-75%, P=.007) and 78% (44-91%, P=.001), respectively. The number of patients needed to treat for 2 years to prevent one primary composite endpoint during this period was 4.3 (95% CI, 3.2-8.8). For end-stage renal disease and all-cause death, the needed numbers of treatment were 5.3 (95% CI, 3.7-15.8) and 6.1 (95% CI, 5.2-11.4), respectively.
Figure 1.
Kaplan-Meier curves for the primary composite endpoint of end-stage renal disease or all-cause death (A), end-stage renal disease (B), and all-cause death (C).
Progression of renal disease
In the usual care group, the serum creatinine concentration (geometric mean [95% CI]) increased from 208.2 (197.8-219.0) μmol/L at baseline to 369.4 (332.0-411.0) μmol/L at study end. In the structured care group, the respective values at baseline and at study end were 206.3 (195.0-218.3) and 309.0 (278.9-342.3) μmol/L. Structured care significantly reduced the rate of change (median [IQR]) in serum creatinine concentration (structured care, 32.5 [159.7]; usual care, 82.4 [240.1] μmol/L/year; P=.032).
Processes-of-care and control of risk factors
Patients in the structured care group had higher frequencies of clinical and laboratory measurements than patients in the usual care group (BP measured ≥4 times/year, 100% vs 72.5%, P<.001; HbA1c ≥2 times/year, 100% vs 86.3%, P=.001; serum lipids ≥1 times/year, 95.0% vs 52.5%, P<.001).
During the study, 74% of patients in the structured care group compared with only 41% in the usual care group required ≥3 BP-lowering drugs (P<.001). Moreover, the use of renin-angiotensin system inhibitor and statin increased significantly during the study compared with baseline in the structured care group. Among the 72 patients who remained in the study at 1 year, 93.1% were prescribed renin-angiotensin system inhibitor compared with 70.8% at baseline (P<.001, McNemar test). The percentage of patients receiving statin was nearly doubled from 37.5% to 70.8% (P<.001). Of the 70 patients who remained in the usual care group at 1 year, the use of statin increased only from 25.7% at baseline to 32.9% (P=.063). The use of renin-angiotensin system inhibitor decreased from 62.9% to 45.7% (P=.012).
At study end, patients in the structured care group had lower systolic (139.7±20.3 vs 148.0±22.5, P=.015) and diastolic BP (67.6±11.8 vs 71.8±11.3, P=.003) compared with patients in the usual care group. Serum concentrations of total and LDL cholesterol were also significantly lowered (structured care vs usual care, 4.4±1.1 vs 5.0±1.4, P=.006 and 2.3±0.9 vs 2.8±1.0, P=.013, respectively). No significant differences in other metabolic indices were observed between groups.
Discussion
We found that structured care provided by a pharmacist-diabetes specialist team emphasizing regular laboratory monitoring, attainment of treatment goals, and patient adherence improved renal outcomes compared with usual care in patients with type 2 diabetic nephropathy and renal impairment. After adjustment for covariates, the risk of the primary composite endpoint was reduced by 60%. The risks of end-stage renal disease and all-cause death were lowered by 55% and 78%, respectively. The renal benefits of structured care were also supported by the reduction in rate of increase of serum creatinine. Given the higher risks for renal deaths in Asian diabetic patients,17 our findings have important implications.
Effective treatments of diabetic nephropathy involve intensive control of multiple risk factors and use of renin-angiotensin system inhibitor.2, 3, 4, 5, 6 Patients are often required to adhere to complex drug regimens and lifestyle changes.18 Moreover, as hyperkalemia and acute renal deterioration may occur with renin-angiotensin system inhibitor,19, 20 close monitoring of serum creatinine and potassium levels is essential. Consequently, the management protocol is complex and often demanding on patients, caregivers, and the healthcare system. Not surprisingly, implementation of these guidelines in routine practice is far from optimal.8
Given the complexity of its treatments, diabetic nephropathy is one of the most ideal chronic conditions for development of a disease management plan.9 To date, no proper disease management programs for this condition have been reported. In studies ever published, the interventions were not described to allow replication in other care settings, the treatment effects were only compared with patients’ historical control, and clinical endpoints such as death or end-stage renal disease were not examined.10, 11 To our knowledge, the present study is the first to demonstrate the effectiveness of a well-defined disease management program on prespecified endpoints of end-stage renal disease and death for patients with type 2 diabetic nephropathy.
Our study was designed to examine the effects of the entire program and has not enabled us to separate the effects of various components. The evidence for benefits of intensive risk factors control is overwhelming.2, 3, 4 Moreover, periodic monitoring of risk factors has facilitated adjustment of regimens.21 Patients receiving structured care had a greater utilization of antihypertensive agents and statin. In randomized trials where tight BP control was the treatment aim, patients often required ≥3 BP-lowering drugs.22, 23 Apart from clinical trial settings, treatments of this intensity are uncommon in routine practice due to presence of multiple patient-24 and physician-related barriers,25 unless a multidisciplinary approach such as ours is adopted.26 We also observed an increase in prescription of renin-angiotensin system inhibitor during the study in the structured care group as opposed to decreased usage in the usual care recipients. In routine practice, clinicians are often discouraged from prescribing or tend to discontinue these agents for fear of the risks of hyperkalemia and acute renal deterioration.19, 20 We believe that renin-angiotensin system inhibitor could have been safely commenced in most patients, including those with more severe renal impairment if a structured approach like ours was used to facilitate close monitoring of serum potassium and creatinine levels as well as correction of laboratory abnormalities.6, 27
The 20% reduction in absolute risk of the primary composite endpoint in our study is higher than that reported in other trials comparing a treatment with placebo in patients with diabetic nephropathy.5, 6 The larger treatment effects observed in our study could be ascribed to the multifactorial nature of our program. Moreover, compared with previous reports,5, 6 our patients were older and had more severe renal impairment. Patients were not excluded for cardiovascular comorbidity alone. As in other intervention trials,23 high-risk patients often benefit more from effective interventions. In this respect, our study results are also more generalizable to the patient population than efficacy trials, which often exclude a substantial portion of patients to whom the intervention is intended to be offered when implemented in routine practice.
We acknowledge that the lack of randomization and blinding could have compromised the study’s internal validity. Due to feasibility constraints, we have adopted a translation research design and used a nonrandomly selected, concurrent control. Nevertheless, several methodological strategies were taken to improve the comparability of study groups. First, we have used the same eligibility criteria to recruit patients into the structured care and usual care groups. All eligible patients were recruited to reduce selection bias. Essentially, no patients refused participation in the structured care program. Second, we have attempted to predefine and adjust for all measurable confounders, although some remained unmeasured, eg, the angiotensin-converting enzyme genotype, which had been related to a faster rate of renal decline.28 Last, the study was designed with predefined primary, secondary, and tertiary endpoints to avoid post hoc and multiple comparisons. Sample size was determined by power calculation. An intention-to-treat approach was used to reduce attrition bias. At baseline, the two groups were relatively balanced. In this respect, several recent meta-analyses suggest that rigorously designed observational studies do not systemically overestimate the effectiveness of interventions for similar conditions.29, 30, 31
Given the growing healthcare burden due to diabetes, more information on the cost-effectiveness of interventions is necessary in order to prioritize the use of healthcare resources. The cost of providing structured care is inevitably higher due to the increased frequency of consultation by pharmacist and diabetes specialist (approximately US $100-120/visit), periodic laboratory tests (US $300-600/year), and intensive drug treatments (US $2340-3740/year). However, it has been estimated that the daily cost of hospitalization is US $380, whereas the annual treatment costs related to peritoneal dialysis and hemodialysis are US $12800 and US $25600, respectively, based on the local public hospital database. In this respect, a cost-effectiveness analysis is currently under way. Given its substantial effect in improving renal outcomes, our structured care model is expected to be cost-effective.
In conclusion, structured care delivered by a pharmacist-diabetes specialist team improved renal outcomes in patients with type 2 diabetic nephropathy and renal impairment compared with usual care. The benefits were attributed to more intensive laboratory monitoring, improved control of risk factors, and more appropriate drug use.
Acknowledgments
We gratefully acknowledge Mr. Albert Cheung, statistician at the Center for Clinical Trial and Epidemiological Research, the Chinese University of Hong Kong, for his expert advice on statistical analyses. Thanks also go to Professor C.E. Mogensen, Medical Department M, Aarhus University Hospital, Denmark, and Dr. Thomas Burke, Worldwide Outcomes Research, Merck & Co., Inc., for their critical appraisal of this article as part of a PhD thesis. Finally, we are indebted to late Professor Julian A.J.H. Critchley for his inspiration and support of this project. The project was supported by a university grant.
References
- . End-stage renal failure in type 2 diabetes (a medical catastrophe of worldwide dimensions) . Am J Kidney Dis . 1999;34:795–808
- . Effect of antihypertensive treatment on kidney function in diabetic nephropathy . Br Med J (Clin Res Ed) . 1987;294:1443–1447
- . Effect of metabolic factors and blood pressure on kidney function in proteinuric type 2 (non-insulin-dependent) diabetic patients . Diabetologia . 1993;36:1051–1056
- . Effect of lipid reduction on the progression of renal disease (a meta analysis) . Kidney Int . 2001;59:260–269
- Renoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes . N Engl J Med . 2001;345:851–860
- Effects of losartan on renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy . N Engl J Med . 2001;345:861–869
- Abysmal prognosis of patients with type 2 diabetes entering dialysis . Nephrol Dial Transplant . 1999;14:129–136
- . The care of patients with diabetic nephropathy (audit, feedback, and improvement) . QJM . 1999;92:443–449
- . Evidence-based disease management . JAMA . 1997;278:1687–1692
- . Can intensive treatment alter the progress of established diabetic nephropathy to end-stage renal failure? . QJM . 1999;92:275–282
- . Diabetic nephropathy (how effective is treatment in clinical practice?) . QJM . 2002;95:41–49
- . Position statement. Diabetic nephropathy . Diabetes Care . 2002;25:s85–s89
- . Nonparametric estimation for incomplete observations . J Am Stat Assoc . 1958;53:457–481
- . Regression models and life-tables . J R Stat Soc Ser B . 1972;34:187–202
- . Calculating the number needed to treat for trials where the outcome is time to an event . Br Med J . 1999;319:1492–1495
- . Introducing Multilevel Modeling . 1st edn.. London, UK: SAGE Publications Ltd; 1998;
- . Mortality and causes of death in the WHO Multinational Study of Vascular Disease in Diabetes . Diabetologia . 2001;44(9 Suppl 2):S14–S21
- . Position statement. Standards of medical care for patients with diabetes mellitus . Diabetes Care . 2002;25:s33–s49
- . Hyperkalemia as a complication of drug therapy . Arch Intern Med . 1987;147:867–869
- . Captopril-induced functional renal insufficiency in patients with bilateral renal-artery stenoses or renal-artery stenosis in a solitary kidney . N Engl J Med . 1983;308:373–376
- . Immediate feedback of HbA1c levels improves glycemic control in type 1 and insulin-treated type 2 diabetic patients . Diabetes Care . 1999;22:1785–1789
- . Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes (UKPDS 38) . Br Med J . 1998;317:703–713
- Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension (principal results of the Hypertension Optimal Treatment (HOT) randomised trial) . Lancet . 1998;351:1755–1762
- . Patient compliance—an overview . J Clin Pharm Ther . 1992;17:283–295
- . Physician-related barriers to the effective management of uncontrolled hypertension . Arch Intern Med . 2002;162:413–420
- . Comparing standard care with a physician and pharmacist team approach for uncontrolled hypertension . J Gen Intern Med . 1998;13:740–745
- . The long-term tolerability of enalapril in hypertensive patients with renal impairment . Nephrol Dial Transplant . 1997;12(Suppl 2):75–81
- Prognostic effect of insertion/deletion polymorphism of the ace gene on renal and cardiovascular clinical outcomes in Chinese patients with type 2 diabetes . Diabetes Care . 2005;28:348–354
- . A comparison of observational studies and randomized, controlled trials . N Engl J Med . 2000;342:1878–1886
- . Randomized, controlled trials, observational studies, and the hierarchy of research designs . N Engl J Med . 2000;342:1887–1892
- . Methods in health services research. Interpreting the evidence (choosing between randomised and non-randomised studies) . Br Med J . 1999;319:312–315
PII: S0002-9343(05)00661-3
doi:10.1016/j.amjmed.2005.07.050
© 2005 Elsevier Inc. All rights reserved.
Volume 118, Issue 12 , Pages 1414.e21-1414.e27, December 2005
