| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Reports |
TIFFANY N. ERICKSON, PHARM.D., BCPS, is Clinical Pharmacist, Department of Pharmacy, Harborview Medical Center (HMC), Seattle, WA; at the time of the study, she was Specialty Resident in Primary Care Practice, Department of Pharmacy, HMC. EMILY BETH DEVINE, PHARM.D., M.B.A., BCPS, FASHP, is Research Assistant Professor, Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy, University of Washington (UW), Seattle. THERESA S. O’ YOUNG, PHARM.D., is Clinical Pharmacist; and LAURA J. HANSON, PHARM.D., BCPS, CDE, is Clinical Pharmacist, Department of Pharmacy, HMC. BENJAMIN FRENCH, M.S., is Ph.D. degree candidate, Department of Biostatistics, UW. CYNTHIA BRENNAN, PHARM.D., M.H.A., is Assistant Director, Ambulatory Pharmacy Services, HMC.
Address correspondence to Dr. Devine at the Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy, University of Washington, Box 357630, Seattle, WA 98195-7630 (bdevine{at}u.washington.edu).
Methods. A retrospective cohort review of patients’ medical records was performed that recorded events occurring between January 1, 2001, and December 31, 2003. The cohort consisted of patients with diabetes in an adult medicine clinic at a county hospital. Patients were included if they were receiving NPH insulin for a minimum of six months and subsequently switched to insulin glargine for a minimum of six months.
Results. The study included 43 patients. There was no significant diffierence in mean glycosylated hemoglobin (HbA1c) between NPH insulin (9.6%) and insulin glargine (9.7%) regimens (p = 0.78, 95% confidence interval, –0.62%, 0.82%). Neither was there a significant difference in the frequency or severity of hypoglycemic episodes between the two treatments. Patients experienced significantly fewer diabetes-associated visits over six months while on insulin glargine. Refill frequency did not differ significantly when patients were receiving NPH insulin versus insulin glargine. When analyzing patient characteristics, those of Hispanic ethnicity experienced HbA1c values significantly higher than white patients. Several characteristics were associated with refill frequency.
Conclusion. The results of our study indicate that both NPH- and glargine-based basal insulin regimens result in similar levels of glycemic control in a medically vulnerable population with diabetes, without significant differences in the number or severity of hypoglycemic episodes or in refill frequency.
Index terms: Diabetes mellitus; Insulin glargine; Insulin human isophane; Insulins; Race; Substitution; Toxicity
Purpose. The purpose of this observational study was to determine if switching from isophane insulin human (NPH) to insulin glargine would improve glycemic control in a medically vulnerable population with uncontrolled diabetes.
According to the Advisory Commission on Consumer Protection and Quality in the Health Care Industry, individuals are considered vulnerable to health care quality problems for many reasons, including financial circumstances, place of residence, health, age, functional status, developmental status, ability to communicate effectively, race, ethnicity, and sex.1 Vulnerable populations can be more prone to health care quality problems, differential experiences in the health care system, or an increased burden of poor health.
Health care quality problems include nonadherence to national standards established for management of chronic conditions, including diabetes. The national standard for optimal blood glucose control is a glycosylated hemoglobin (HbA1c) of 
7%.2 To achieve this level of control, it is often necessary to begin insulin therapy.3 An optimal exogenous insulin regimen mimics physiological insulin secretion in the healthy individual. This includes maintaining basal insulin as provided by long-acting insulins and the use of short-acting insulin at mealtimes. While the goal is usually to use multiple injections of insulin to control blood sugar at all times of the day, patient characteristics such as poor comprehension, fear of injections, and lack of follow-up are often barriers to rapid insulin adjustment, the initiation of which can be done in a stepwise manner, beginning simply with basal insulin given once or twice daily.
Although isophane insulin human (NPH) or long-acting insulin glargine (rDNA orgin) (Lantus, Sanofi-Aventis) is often used as the basal insulin of choice, there is no consistent evidence that either of these insulins is more effective in decreasing baseline HbA1c in adults with type 1 or type 2 diabetes mellitus, regardless of the use of simultaneous antidiabetic therapies.4–13 As a basal insulin, glargine has the advantage of once-daily dosing, and evidence suggests it may be associated with a lower incidence of hypoglycemia when compared to NPH insulin.4,5,7–13 However, patients with medical vulnerabilities who would most benefit from these advantages, including those who have a history of substance abuse, mental illness, or poor comprehension, often are excluded from controlled clinical trials, either by study design or loss to follow-up.5,8,10
Studies have shown that medication adherence is associated with glycemic control, with increased adherence leading to decreases in HbA1c.14–18 This has also been seen in an indigent population.19 These studies have noted several patient factors associated with poor adherence, including the patient’s emotional wellbeing, comprehension, economic status, medical complexity, and ethnicity or race.17,18 Based on these findings, it is thought that once-daily basal insulin would give higher adherence rates than twice-daily basal insulin, especially in a population at risk for poor adherence. There is little information on diabetes-related outcomes and adherence to different basal insulin regimens in a medically vulnerable population, and even less is known about whether individual factors affect the outcomes.
Finally, drug therapy and associated health care costs are of concern—especially at a county hospital that provides a large proportion of charity care to an indigent population. One online retailer of prescription drugs lists the price of insulin glargine as almost twice as expensive per milliliter as NPH insulin.20
As the county hospital for a large metropolitan area, our institution cares for a medically vulnerable population that includes over 750 diabetic patients and in which diabetes is the second most commonly diagnosed chronic disease. A 2001 review of HbA1c values in our adult medicine clinic revealed that 60% of patients had a value above 7%. To give providers more options for improving glycemic control, insulin glargine was added to our formulary in 2001, shortly after it was marketed. Subsequently, its use steadily increased, enough to surpass that of NPH insulin. Better adherence to a once-daily basal insulin regimen and the hope of a reduction in hypoglycemia in a vulnerable population are believed to be the reasons for switching from NPH insulin to insulin glargine. However, our institution’s purchase price for insulin glargine is almost three times that of NPH insulin per milliliter. Without convincing data that insulin glargine provides superior efficacy when compared to NPH insulin, and with minimal data on outcomes in vulnerable populations, it is unknown if the increased drug cost would offset other expenditures associated with diabetes care, including the frequency of health care visits for uncontrolled diabetes.
The purpose of this retrospective cohort study was to explore whether switching from NPH insulin to insulin glargine in a medically vulnerable group of patients with uncontrolled diabetes would improve the following outcomes: level of glycemic control, frequency and severity of hypoglycemic episodes, frequency and type of diabetes-associated visits, and medication adherence as measured by prescription refill frequency. A secondary interest was to explore whether specific patient characteristics could influence either the level of glycemic control or basal insulin prescription refill frequency.
 |
Methods |
|---|
 Top Methods ResultsDiscussionConclusionReferences |
|---|
Design. This was a retrospective cohort review of patients’ medical records documenting outcomes of interest occurring between January 1, 2001, and December 31, 2003. The cohort consisted of Adult Medicine Clinic patients at our institution. Patients were included if they had a diagnosis of diabetes, received NPH insulin for a minimum of six months, and were subsequently switched to insulin glargine for a minimum of six months during the study period. Patients were identified from centralized pharmacy distribution system records. The study was approved by the University of Washington Human Subjects Review Committee, under a waiver of consent.
The primary outcome measure was the HbA1c value (percent). Secondary outcomes included the frequency and severity of hypoglycemic episodes, the frequency of diabetesassociated clinic visits, emergency department (ED) visits or inpatient admissions, and basal insulin refill frequency, comparing NPH-based to insulin glargine-based regimens. To determine if other factors may have affected the outcomes of interest, the use and dosages of basal and prandial insulin and the use of oral antidiabetic medications were measured. Patient characteristics associated with vulnerability, including sex, ethnicity and race, the presence and number of diagnoses of mental illness, the presence of substance abuse, primary language, prescription insurance type, and place of residence, were identified and evaluated for their effect on the primary outcome and on refill frequency.
Variable identification. The HbA1c value was identified as the laboratory value measured at the time closest to the end of each six-month basal insulin interval, as long as it was at least three months after initiation of the basal insulin regimen. Hypoglycemic episodes and diabetes-associated visits were included if they occurred at any time during the six months of each type of insulin therapy. These episodes and visits were measured for frequency and severity. We used weighted severity descriptors for these two outcomes because differences in severity have different levels of clinical and economic significance, and because low numbers of severe episodes were expected. Less severe hypoglycemic episodes (severity level 1) were defined as a laboratory value indicating a blood glucose concentration of less than 70 mg/dL, or a chart note indicating the same, based on a patient’s self-monitoring. Hypoglycemic episodes of greater severity (severity level 2) were included if the pharmacy record indicated a prescription for glucagon was filled or if the patient had an ED visit with the primary diagnosis of hypoglycemia, based on International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) coding. Diabetes-associated visits were identified through a review of billed-visit history, where the primary billing diagnosis included diabetes, again based on ICD-9-CM coding. These were rated for severity level with an ambulatory clinic visit coded with a severity level of 1; an ED visit, level 2; and an inpatient admission, level 3.
Prescription refill information was obtained from the centralized pharmacy dispensing system and was determined by dividing the number of basal insulin prescription refills obtained by the number of prescription refills expected and expressed as a percentage. This method of determining refills to estimate adherence has been used elsewhere and has been validated.21,22 The dosages of basal and prandial insulin, if applicable, and the types and dosages of oral antidiabetic medications were determined through a review of chart notes or pharmacy refill records; dosages documented closest to the end of each six-month period were noted. Oral antidiabetic medications were grouped according to therapeutic class based on local usage: metformin, sulfonylureas, and thiazolidinediones.
Patient demographics and characteristics were recorded from the medical chart. The type of diabetes was classified as type 1, type 2, or due to secondary causes such as pancreatitis or prolonged corticosteroid use. A diagnosis of a mental health disorder was included if the patient had an ambulatory visit with the primary ICD-9-CM diagnosis of depression, bipolar disorder, schizophrenia, anxiety, panic, personality disorder, posttraumatic stress disorder, or psychosis not otherwise specified. A history of substance abuse was included if the primary ICD-9-CM diagnosis for a visit included substance-abuse disorder, i.v. drug abuse, or alcohol use or if any of these were noted in a chart note. Residence was classified as stable if it was at one site for the duration of the study, or as unstable if the patient changed residences or none was identified during the study. Prescription insurance was determined from the pharmacy dispensing system.
Statistical analyses. Univariate analyses consisted of calculating percents or means and standard deviations of patients’ demographic characteristics. Bivariate analyses, comparing the primary and secondary outcomes and antidiabetic medication use when patients were receiving NPH insulin versus insulin glargine, were conducted using the following regression techniques: linear regression to analyze HbA1c values and the number of units per basal and prandial insulin regimens; logistic regression to analyze oral antidiabetic medication use; and Poisson or negative binomial regression to analyze the numbers of injections of basal and prandial insulin, hypoglycemic episodes, visits, and prescription refills. Negative binomial regression was used in place of Poisson regression when overdispersion (i.e., variance greater than the mean) was present.23 For ease of discussion, the results of the Poisson and negative binomial analyses are presented as incidence rate ratios, which use 1.0 instead of zero as a reference point for interpreting confidence intervals (CIs).23 Multiple linear regression was used to determine the effect of patients’ characteristics on the primary outcome. Negative binomial regression was used to evaluate the effect of characteristics of medical vulnerability (age, sex, ethnicity and race, mental illness, substance-abuse disorders, primary language, prescription coverage, type of residence) on basal insulin prescription refill frequency. Categorical variables were coded such that the largest group served as the referent. In the model evaluating the primary outcome, prandial insulin and oral antidiabetic medications were included as potential confounders.
Because each patient was placed under observation twice and served as his or her own control, within-patient clustering was taken into account in each analysis, using the Huber–White (sandwich) estimator. 23 Each variable was analyzed for its significance as an interaction or a confounding variable. All tests were two-tailed, using an 
level of 0.05 and 95% CIs. All statistical analyses were completed using Stata, version 8SE (Stata Corp., College Station, TX).
To evaluate whether the mean HbA1c of patients in the study differed from that of the overall population seen in the Adult Medicine Clinic, the mean HbA1c of the study patients was compared to that of all insulin-dependent diabetics seen in the clinic in 2003.
|
Results |
|---|
|
TopMethodsResultsDiscussionConclusionReferences |
|---|
A summary of the patients’ characteristics appears in Table 1
. Results from the primary and secondary outcomes analyses are reported in Table 2. The mean HbA1c was 9.6% and 9.7% for patients receiving NPH and insulin glargine, respectively. This difference was not statistically significant. Neither was it significantly different from the mean HbA1c value of insulin-dependent patients seen in the Adult Medicine Clinic in 2003 (n = 384, 9%, p = 0.10). Twenty-three patients had an increase in HbA1c and 20 patients had a decrease in their HbA1c when on insulin glargine. There was no difference in the severity-weighted frequency of hypoglycemic episodes during treatment with either insulin. The number of clinic visits, number of ED visits, and mean number of total visits (weighted by severity) were significantly lower when patients received insulin glargine, but the difference in diabetes-associated inpatient admissions was not significant. Although not statistically significant, a lower proportion of basal insulin prescription refills was obtained for basal insulin when patients were on NPH insulin versus insulin glargine—75% versus 86%.
|
|
summarizes the medication-use characteristics of patients, comparing NPH insulin to insulin glargine treatment. The difference in the mean number of units administered per dose of basal insulin approached significance (p = 0.06), with patients injecting about 12% more total daily units of basal insulin when on NPH insulin versus insulin glargine. As expected, there was a significant difference between the number of injections per day, with insulin glargine being dosed with one half the frequency of NPH insulin. Table 3 lists the number of patients receiving prandial insulin, the number of units and injections per day, and whether or not patients were taking one or more oral antidiabetic medications. There were no significant differences in these categories when comparing NPH insulin and insulin glargine. Finally, a description of the proportion of patients receiving oral antidiabetic agents, by therapeutic class, is also found in Table 3.
|
summarizes the results obtained from the multiple linear regression, evaluating the effect of patient characteristics on HbA1c levels, while controlling for treatment regimen. The model contains only main effects, as no important interactions were identified. Both prandial insulin and oral antidiabetic medications were retained in the model, as when these were included, the difference in mean HbA1c obtained with the NPH insulin versus glargine regimens decreased, from a mean difference of 0.1% (found in the bivariate analysis) to a mean difference of 0.12% in the regression model. Of significance was that Hispanics had HbA1c values that were 4.14% greater than those of whites.
|
displays the results of the regression model used to explore the effect of characteristics that indicate medical vulnerability based on the proportion of prescription refills, while controlling for treatment regimen. Four significant results were found. Patients who obtained a significantly lower proportion of prescription refills included females (versus males; p < 0.05), those with no insurance or with the low-income allowance (versus Medicaid; p < 0.05), Native Americans (versus white; p < 0.01), and those with one mental health diagnoses (versus two diagnoses; p < 0.01). Although there were no other significant differences based on patient characteristics, the direction of the rate ratios is noteworthy. Other characteristics that led to a decreased proportion of refills obtained included being black or Hispanic versus white, having one diagnosis of substance abuse versus none, being non-English speaking versus English speaking, and having no insurance or solely a low-income allowance or Medicare versus Medicaid. Characteristics that led to an increased proportion of refills obtained included each year of increasing age, being Asian versus white, having no mental health disorders versus two, having two or more diagnoses of substance abuse versus none, having the Basic Health Plan or private insurance versus Medicaid, and having a stable residence versus unstable residence.
|
|
Discussion |
|---|
|
TopMethodsResultsDiscussionConclusionReferences |
|---|
The patient demographics reflect the medically vulnerable population served at our institution. In our study, we were unable to detect a difference between the mean HbA1c values or a difference in the frequency and severity of hypoglycemic episodes after six months of treatment with NPH insulin versus insulin glargine. When receiving insulin glargine, almost equal numbers of patients experienced an increase in HbA1c as a decrease. Further, the standard deviations and ranges for HbA1c during treatment with NPH insulin or insulin glargine are similarly distributed, suggesting that outliers were unlikely to have affected the mean. It is more likely that there truly was no difference, as patients had an equal chance that HbA1c would increase or decrease when switching to insulin glargine. Furthermore, there was no difference in the number of patients who reached the goal HbA1c, 7%, when on the two insulin regimens. Separately, the fact that we were unable to see a difference in hypoglycemic episodes is likely because our population had such poor, underlying glycemic control. With tighter control, it is possible that insulin glargine would be useful in decreasing the incidence of hypoglycemia in a vulnerable population. Importantly, the average weight of the patients was 92.7 kg, reflecting suboptimal weight control and diabetes control.
Overall, patients used a marginally lower number of units of insulin glargine than NPH insulin. This is to be expected, as the recommended starting dosage of insulin glargine is 80% of the NPH insulin dosage, and the dosage of insulin glargine may have been adjusted upward during the six-month treatment period. Neither the use nor dosage of prandial insulin, or the use of oral antidiabetic medications differed between the two groups, so it is unlikely these are affecting the outcomes examined.
When comparing HbA1 c values between the two basal insulin regimens, the results of our study were similar to findings previously reported, although those trials did not specifically examine a vulnerable population.4–13 In contrast, a recent, similarly designed retrospective study of patients receiving NPH insulin with inadequate glycemic control who switched to insulin glargine showed a significant decrease in HbA1c values at the end of one year of treatment with insulin glargine, when compared with values while receiving treatment with NPH insulin.24 However, the authors noted that other antidiabetic medications and prandial insulin were adjusted throughout the study and this could have been the reason for the improvement in HbA1c at the end of the study. These patients also had better overall glycemic control than our population and were receiving care from endocrinologists at private practices sites; thus, these patients are unlikely to have the same characteristics as our indigent population.
In a separate study, investigators examined the change in HbA1c values when patients with type 1 diabetes mellitus treated with once-daily NPH insulin were switched to insulin glargine.25 They found a significant reduction of 0.5% in HbA1c values at one year. This study has several limitations as well. It was unblinded and not randomized, and patients received once-daily NPH insulin, which is lower than usually recommended, before being switched to once-daily insulin glargine. The dosage of insulin glargine was intentionally adjusted upward, with the eventual goal of twice-daily glargine dosing. This study included only patients with type 1 diabetes.
The lack of a significant effect of changing the long-acting insulin regimen on HbA1c values observed in our study is similar to findings that have been reported when comparing the use of short-acting insulins. A systematic review of short-acting insulin analogues and their impact on HbA1c concluded that there may be a minor benefit in patients with type 1 diabetes mellitus; however, in the remaining population, there was no clear benefit of using insulin analogues when compared to using regular insulin.26
Interestingly, the use of insulin glargine in our study was associated with fewer diabetes-associated visits. We could not determine the reasons for this finding; however, the decrease in diabetes-associated visits may offset the increased cost of insulin glargine. At our institution, if a patient were to receive one vial of insulin glargine per month for six months, this would approximate the cost of one routine follow-up clinic visit lasting approximately 20 minutes.
Recently, Sokol et al.27 examined the effect of medication adherence on health care use and cost associated with diabetes care. The investigators found that a high level (80–100%) of adherence was associated with lower overall diabetes-related medical costs, and higher medication costs were more than offset by medical cost reductions. The patients in our study reached 80% adherence, on average, when receiving insulin glargine. The study of Sokol et al. is limited in that it was observational and included patients who were enrolled for two years in a large manufacturing employer’s health plan. In addition, the population was younger than 65 years, which limits generalizability, especially to the indigent, elderly, and unstable populations.
The authors of a separate evaluation concluded that insulin glargine is cost-effective in patients with type 1 diabetes mellitus and borders on cost-effectiveness in patients with type 2 diabetes mellitus.28 They attributed the difference in cost solely to the number of hypoglycemic episodes that were avoided by using insulin glargine. However, they noted that because there was no improvement in long-term glycemic control, it is unlikely that the regular use of insulin glargine will reduce the incidence and costs of long-term complications associated with diabetes. Finally, a comparison of insulin lispro, a short-acting insulin analogue, versus regular insulin resulted in higher diabetes and pharmacy-related costs but lower nondiabetes medical costs, leading to similar total costs overall.29 These studies demonstrate that, to date, there is no convincing or prospective evidence that the use of insulin glargine as basal insulin, regardless of adherence rates, is associated with lower health care costs when compared with NPH insulin, and costs should not be considered the sole reason to switch patients to insulin glargine.
Although the difference between treatment regimens was not statistically significant, in our study refill frequency increased when patients were treated with insulin glargine. Adherence may be improved by using a once-daily versus twice-daily basal insulin; however, as the insulin regimen continues to be adjusted, the total number of injections per day may counter this benefit, as prandial insulin cannot be mixed with insulin glargine and administered as a single injection, but can be with NPH insulin. Regardless, in our population, there was no difference in outcomes when adherence was improved. This suggests that much more contributes to outcomes in patients with uncontrolled diabetes than the use of and adherence to basal insulin.
Our evaluation of patient characteristics did not reveal any significant differences in HbA1c when switching from NPH to insulin glargine. Regardless of the characteristic that makes a patient vulnerable to poor glycemic control, using insulin glargine does not appear to improve HbA1c values when compared with NPH.
The significant difference in HbA1c values between Hispanics and whites observed in our study has been reported elsewhere.30–33
The results of the regression model exploring patients’ characteristics on refill frequency are revealing. Patients who are female, black, Hispanic, or of Native American heritage obtain fewer refills than do those who are white or Asian. It is not intuitively consistent that patients with one mental health diagnosis would obtain fewer refills than those with two or more diagnoses of mental illness. However, this finding can be explained in that, at our institution, patients who are more disadvantaged are often given additional assistance; for example, they are supported by case workers who assist with diseasestate management and are given prefilled medication organizers. A similar argument can be made for patients with substance-abuse disorders, those with one diagnosis receiving fewer refills, and those with two or more diagnoses receiving more refills than those without a disorder. Logical, too, is that those without prescription insurance coverage, or those with Medicare, obtained fewer refills than those with Medicaid; and that those with the state of Washington’s Basic Health Plan or private insurance obtained more refills than those with Medicaid. It would be interesting to make these same comparisons now that Medicare Part D is in effect. Finally, it is logical that patients without stable home environments obtained fewer refills than those with stable home environments.
Our observational study was exploratory and has several limitations. The possibility of bias exists because our patients were not randomized to a treatment course (i.e., all patients first received NPH insulin, then insulin glargine). However, this bias may be small because the six-month interval between HbA1c measurements provided an adequate washout period, preventing any residual effect of NPH insulin on outcomes while patients were receiving insulin glargine. Separately, the fact that one patient receiving NPH insulin and eight patients receiving insulin glargine were excluded because they did not remain on the basal insulin treatment for six months because of inadequate HbA1c control limits generalizability. The inclusion of these patients may have affected the measured outcomes. However, because of the inability to reach HbA1c goals, it is unlikely that including these patients would have resulted in significantly improved glycemic control when patients were receiving insulin glargine. We do know that the HbA1c values of the patients in the study were similar to those seen in the clinic, overall.
Our sample size was small. This is partially due to our vulnerable population being lost to follow-up and lacking complete chart data. Because of the small numbers, we did not have adequate power to detect a statistically significant difference of 0.1% in mean HbA1c values between the treatment regimens. This does not diminish a related question, which is whether a difference of 0.1% is clinically significant. To put the results into perspective, a sample size of 43 patients did provide 80% power, which allowed us to detect an HbA1c difference of 1%. Caution is warranted in interpreting the effect of patient characteristics on the outcomes, as the numbers in some of these strata were small.
Reporting bias may have been introduced by patients’ self-reports of hypoglycemic episodes. Misclassification bias may have occurred in that a patient was counted as having a diabetes-associated visit only if diabetes was coded as the primary diagnosis for that visit. There may have been additional visits where diabetes was addressed and managed but not counted in this study. In the chart review, other factors associated with diabetes care were not examined. These include dietary and activity patterns as well as refill frequency of oral antidiabetic medications, which could have an effect on glycemic control. Patient satisfaction was also not measured, and this could have an impact on adherence rates and on health care visit frequency. Our study was limited to patients whose prescriptions were filled at our institution; however, this is unlikely to be an issue, as the incentive to obtain medications—provision without upfront payment—is sufficiently large to have created a closed system of health care. Finally, we recognize that although we measured the number of refills of basal insulin obtained by our patients, this does not necessarily correspond with their adherence to injecting the insulin.
When patients have uncontrolled diabetes during treatment with basal insulin, there are options for improving glycemic control other than switching basal insulin type. The dosage of basal insulin used by patients in our study was relatively low (0.6 unit/kg/day) for a population that had, on average, a baseline HbA1c of 9.6% and weighed 92.7 kg. It has been suggested that in order to achieve an HbA1c of <7%, the daily dosage of basal insulin can be adjusted upward to 1 unit/kg.33 It is also possible that better glycemic control may have been obtained by increasing the use of prandial insulin, as only two thirds of our population was using prandial insulin. Attempts at increasing adherence with the initial basal insulin regimen could be made. Addressing diet and activity patterns and weight loss would likely safely and inexpensively improve glycemic control in a vulnerable population.
The decreased frequency and severity of clinic visits prompt the question of costing each regimen; this should be further explored. Conducting a more formal study to confirm these results, to cost these differences, and to examine outcomes in vulnerable populations would be beneficial.
|
Conclusion |
|---|
|
TopMethodsResultsDiscussionConclusionReferences |
|---|
|
References |
|---|
|
TopMethodsResultsDiscussionConclusionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
