a) The authors write the following about this analysis in their paper
“There was a statistically significant change in the proportion of patients whose HbA1c
level was > 7.0% at baseline compared with follow up; 8 of 30 patients (26.67%) moved from the > 7.0% to
c2, 13.42;
P > 7.0% category.”
Their conclusion is INCORRECT! It gets much worse when you consider other outcomes, particularly obesity and BMI where the others appear to have extremely strong evidence of improvement, when in fact they DO NOT even come close to having significant results!!! Why would I say this?
b) Conduct an appropriate analysis for this outcome and state the correct conclusion using a = .05 level of significance.
·
c) Do the same for the obesity result (2nd
to the last row of the table) and compare contrast your results with those of the authors’.
M M t m h a a a t f m s g i m e c u H 0 Clinical Therapeutics/Volume 35, Number 1, 2013 Short and Long-Term Outcomes from a Multisession Diabetes Education Program Targeting Low-Income Minority Patients: A Six-Month Follow Up John G. Ryan, DrPH1; Terri Jennings, PhD2; Isabel Vittoria, MS, LMHC1, and ark Fedders, MS1 1Department of Family Medicine and Community Health, University of Miami Miller School of Medicine, iami, Florida; and 2Private Practice, Davie, Florida t i c ABSTRACT Background: A diabetes self-management educa- ion (DSME) program was offered to patients at a pri- ary care clinic serving low-income people. Objectives: The purpose of the analyses presented ere was to understand the feasibility of the program nd effectiveness of the intervention. Methods: The program was facilitated by a nurse nd licensed dietician. Data were collected at baseline, fter each class, and after 6 months. Patients were in- erviewed to identify diabetes self-care behaviors be- ore the first class, after the fourth class, and at 6 onths. Knowledge related to content areas was mea- ured before and after each class. Glycosylated hemo- lobin (HbA1c), blood pressure, weight, and body mass ndex (BMI) were collected at baseline and after 6 onths. Medical records were reviewed for LDL lev- ls, co-morbidity, and diabetes management. Frequen- ies, �2 and t tests, and repeated measures t tests were sed to analyze data. Results: Patients were mostly non-Hispanic black or ispanic (93.1%); mean BMI was 34.89 kg/m2. About one-half (41.95%) completed the program. Significant improvements were observed for knowledge related to each of the 4 content areas: diet (P � 0.001), diabetes management (P � 0.003), monitoring blood glucose (P � 0.001), and preventing complications (P � .001). Among long-term outcomes, mean HbA1c was significantly reduced (0.82%), from 8.60% to 7.78% (P � 0.007), with 26.67% of patients reducing HbA1c from �7.0% at baseline to �7% at follow up (P � 0.001). Patients demonstrated a significant improvement in readiness to improve dietary behaviors (P � 0.016). Conclusions: Outcomes suggested that minority pa- tients with a high risk for poor diabetes outcomes might be retained in a multisession DSME program and benefit from increasing knowledge of diabetes con- January 2013 tent. Further evaluation is necessary to determine the cost-effectiveness of this intervention. (Clin Ther. 2013;35:A43–A53) © 2013 Elsevier HS Journals, Inc. All rights reserved. Key words: Diabetes, public health, self-efficacy, education, African American INTRODUCTION Minorities in the United States experience dispropor- tionately worse diabetes outcomes than non-Hispanic whites; however, African Americans have a greater risk for diabetes,1 are more likely than other groups to have worse diabetes outcomes,2,3 and are more likely to have poor measures of diabetes control.4–7 An impor- ant consideration for managing diabetes and prevent- ng complications is that �95% of diabetes care is ac- omplished by the patient.8 Research on African Americans with type 2 diabetes suggested that internal factors, including lack of self- control and memory failure, represented barriers pre- venting them from engaging in diabetes self-manage- ment.9 Poor memory and other cognitive challenges have been associated with low health literacy.10,11 These intrinsic characteristics, which increase the risk for poor diabetes outcomes among low-income Afri- can Americans, suggest that this population has a unique need for interventions that have limited de- mands on cognitive skills and that are designed to in- crease intrinsic factors for effectively engaging in diabetes self-management, contributing to improved Accepted for publication December 13, 2012. http://dx.doi.org/10.1016/j.clinthera.2012.12.007 0149-2918/$ - see front matter © 2013 Elsevier HS Journals, Inc. All rights reserved. A43 http://dx.doi.org/10.1016/j.clinthera.2012.12.007 p b v p a o r c t t p w c c f p i f H t H g c m c o n l b a c m a t f t r m C h u s Clinical Therapeutics diabetes outcomes, such as knowledge, understanding, and self-efficacy. Diabetes self-management education (DSME) is rec- ommended by the American Diabetes Association (ADA) for all people with diabetes. The ADA recom- mends 9 content areas for DSME curricula,12 including diabetes treatment outcomes, nutritional management, physical activity, medication, self-management, acute and chronic problems, and personal strategies to ad- dress psychosocial issues and behavioral change. Self- management education teaches problem solving through enhancing self-efficacy.13 DSME produces ositive long-term outcomes, including reduced mor- idity and improved quality of life.14 Nevertheless, DSME is frequently not accessible to patients with the highest risk for poor diabetes outcomes, which in- cludes low-income, African American, and other mi- nority patients with diabetes and no health insurance or Medicaid.15 There is little available evidence to pro- ide guidance with regard to whether such an intense rogram is feasible, given the barriers that are present mong this population. In an effort to increase DSME access to patients in ur target population, characterized as low-income Af- ican Americans and Latinos who were followed in a ommunity-based primary care clinic affiliated with a ertiary care hospital in an urban environment, we in- roduced an intensive yet cognitively and culturally ap- ropriate DSME curriculum. The program was aligned ith ADA recommendations and facilitated by health are providers in conjunction with other diabetes- entric, patient-centered activities. We report on the easibility of implementing such a program as well as rogram outcomes. METHODS A curriculum in English and Spanish was developed from the 9 ADA-recommended topics.12 Other tools ncluded data collection tools adapted with permission rom a best practices clinic, Gateway Community ealth Center, in Laredo, Texas, and patient educa- ion materials in English and Spanish obtained from ealthy-Interactions. A 4-week, 4-class, 12-hour pro- ram was designed with 1 class each week, with each lass representing a content theme: diet, diabetes anagement, blood glucose monitoring, and compli- ations. Diabetes support activities were also devel- ped and made available, representing a longitudi- al diabetes support program framed by social h A44 earning theory.16 –18 The same staff, fluent in English and Spanish, facilitated both English and Spanish classes, thereby enhancing fidelity to the curriculum. Patient learning was encouraged through a collabora- tive approach,13 including use of educational materials and provocative questioning. Referrals for participa- tion in the intervention were received from 4 sources: primary care physicians, an advanced registered nurse practitioner (ARNP), a licensed dietician (LD), and an exercise physiologist. Processes regarding referrals, re- cruitment, presentation as scheduled, and program completion were tracked to support our inquiry re- garding feasibility. Patient demographic characteristics, clinical mea- sures, self-reported behaviors, and availability of sup- port to manage stress were collected at baseline. De- mographic characteristics included birth date, gender, race/ethnicity, preferred language, home language, ed- ucation, and income. Short-term outcomes included changes in knowledge related to each of the 4 content areas covered: nutrition, diabetes management, moni- toring blood glucose levels, and avoiding diabetes com- plications. Change in knowledge was measured with class-specific tests administered before and after each of the classes. Knowledge scores were calculated as the number of correct responses divided by the number of responses. Changes in knowledge were calculated as post-score minus pre-score for patients who took both pre- and post-tests for a given session. Long-term outcomes included changes in clinical, physical, and patient-reported outcomes. Clinical and physical measures collected at baseline and 6-month follow up included glycosylated hemoglobin (HbA1c), lood pressure, fasting or random blood sugar levels, nd weight and height. Body mass index (BMI) was alculated. LDL cholesterol levels, type of diabetes, co- orbidity, and manner in which diabetes was man- ged was obtained from patient medical records. Pa- ient-reported measures collected at baseline and ollow up included changes in general health, medica- ion adherence, readiness to change eating habits, and eadiness to change exercise habits. Patient-reported easures were collected with a tool used at Gateway ommunity Health Center, which was adapted for ealth literacy. Medication adherence was measured sing the 4-item Morisky Medication Adherence Mea- ure.19 Self-assessed readiness-to-change and general ealth measures used 5-point Likert scales. Volume 35 Number 1 s c t a c c g t F F p c T w c i m d c t a t a � H f l w 1 0 s a P m J.G. Ryan et al. Data from patients for the 32-month period from March 2009, when the program was initiated, through October 2011, were used in these analyses. Statistical analyses were completed with SPSS (version 15.0; SPSS Inc., Chicago, Illinois). Continuous variables were ex- pressed as the mean (SD) and range. Continuous vari- ables were compared using repeated measures t test. Discrete variables were expressed as counts and per- centages, and the �2 statistic was used to compare pro- portions. Predictors of intervention completion status were examined using logistic regression. All statistics were 2-tailed, and P values � 0.05 were considered tatistically significant. Program costs were compared with patient care osts due to diabetes complications, to evaluate poten- ial costs to the health care system for this educational nd public health intervention. Program costs were in- urred from compensating clinical staff to facilitate lasses and administrative staff to support the pro- ram. Clinical staff included an ARNP (0.012 FTE [full ime equivalent]), LD (0.16 FTE), and LPN (0.125 TE). Support staff included a research assistant (0.50 TE), patient access representative (0.19 FTE), and rogram manager (0.05) FTE. Costs from diabetes omplications were based on 2005 estimates from owers Watson HR Consulting.20 These estimates ere based on potential cost savings, given annual in- idence and cost of event, of achieving the ADA quality ndicators of HbA1c �7%, LDL cholesterol �130 mg/dL and �100 mg/dL, and blood pressure �140/90 m Hg and �130/80 mm Hg. Program costs were erived from estimating the efforts contributed by ourse facilitators and program staff, and calculating otals of associated salaries and benefits. These evalu- tions did not include a cost-effectiveness approach. RESULTS A total of 543 patients were referred from March 2009 to October 2011; 542 were contacted; and 226 were recruited by scheduling them for the first class in the series (43.3% recruitment rate). One hundred seventy- four enrolled by presenting to the first class (77%). Eleven patients became ineligible during the course of the program, and 73 completed the program (41.95% completion rate). The Figure illustrates patient flow. Most patients were non-Hispanic black (57.5%); 35.6% were Hispanic. The majority of patients were uninsured (35.6%). Physical measures at baseline were as follows: weight, 219.04 (48.41) lbs and BMI, 34.89 � January 2013 (8.03) kg/m2; 8.2% were in the normal BMI range, 17.8% were in the overweight BMI