Sample

The data were collected as part of a larger study on patients in family and internal medicine in the Meharry Medical Practice Plan (MMPP) primary care clinic, at Meharry Medical College in Nashville, TN, which serves the local low-income, largely Medicaid-eligible population. The study enrolled new and existing patients who presented for primary care. Convenience sampling was used since logistics precluded random selection. The original sample was structured to yield equal proportions of persons aged 18-49, 50-64 and 65+ years. All participants were asked to complete a structured interview lasting about one hour. A sub-sample of African-American participants aged 40 years and older (n=303) was used for the present analysis. This African-American subsample was largely low-income, female, and served by TennCare (Tennessee’s Medicaid-waiver program) (Table 1). Over one-fourth of the sample (n=88, 29%) reported that they had diabetes. Multivariate analyses were estimated, using listwise deletion, on the 298 participants with complete data for the variables included in models.

Measures

Table 1. Background Characteristics and Utilization Levels of African-American Participants, Age 40 and Over (Meharry Medical Practice Plan, Nashville, TN; 2001)

No Diabetes Diabetes P
n 215 88
Age (mean number of years)

59.8

62.0

0.13
Gender Percent male Percent female 30.7 69.3 33.0 67.0 0.701
Monthly Household Income Percent under $750 Percent $751-$1,250 Percent Over $1,250 55.7 22.4 21.9 69.4 20.0 10.6 0.045
EducationPercent less than high school Percent high-school graduate Percent postsecondary degree 42.8 41.9 15.3 57.9 27.3 14.8 0.037
InsurancePercent TennCare (Medicaid)

59.3

63.2

0.533
Comorbid Chronic Conditions% with hypertension and/or cardiovascular disease

73.5

86.4

0.015
Physician Visits (last six months)

3.2

3.5

0.436
ER Visits (last 12 months)

0.7

0.7

0.902
Inpatient Days (last 12 months)

1.7

2.2

0.418

The presence or absence of diabetes (type 1 or type 2) and comorbid chronic conditions (hypertension and/or cardiovascular disease) were based on self-reports by the study participants during the structured interview. A condition was counted as present (=1) if the respondent reported having suffered from this condition during the past 12 months and having received treatment for it either currently or in the past. Since the data were collected as part of a larger study, it was not possible to confirm diagnoses with chart reviews. However, we feel confident that patients are very likely to know if they have been diagnosed and treated for diabetes, since this condition requires a relatively intensive treatment regimen. While self-reports of hypertension and cardiovascular disease could be less accurate, it is most likely that participants would err on the side of not reporting one of these conditions when they actually have been diagnosed and treated. We feel that this possibility does not seriously affect our analysis, since this type of error would bias the analysis toward the null hypothesis of no effect and lead to more conservative estimates. Two measures of depressive symptoms were used:

a) The presence of major depressive symptoms was measured using a nine-item symptom checklist from the mood module of the PRIME-MD Clinician Evaluation Guide. While the checklist does not provide a medical diagnosis of depression, it permits the classification of respondents who exhibit symptoms of major depressive disorder based on DSM-IV criteria and would likely receive a diagnosis if evaluated by a clinician. Symptoms of major depression are considered present (depressecNl) if respondents indicate that they have experienced five or more of the nine symptoms a lot during the past two weeks (with one of the symptoms being either “little interest of pleasure in doing things” or “feeling down, depressed, or hopeless”).

b) The Center for Epidemiologic Studies Depression Scale (CES-D) was also used as a general measure of severity of depressive symptomology. The CES-D is a 20-item scale that measures the frequency of symptoms during the preceding week (each item ranges from 0=none to 3=5-7 days). The 20 items are summed (with positive items reversed), resulting in a total score ranging from 0 to 60. Higher scores indicate greater number and frequency of depressive symptoms. A standard cut-off score of
16 was also used to categorize participants with elevated depressive symptoms, as an additional measure of the presence of depressive symptoms.

Levels of healthcare utilization were also based on self-reports from the structured interviews. The three types of healthcare utilization examined were: 1) physician visits—the number of visits to a physician in the last six months; 2) ER visits—the number of ER visits due to a medical problem during the last 12 months; and 3) inpatient days—the number of inpatient hospital days during the last 12 months. Self-report measures of utilization of services were preferred rather than using clinic or hospital records, since the patients may have used physician, ER, or inpatients services at other locations as well. cheap cialis canadian pharmacy

The following demographic characteristics were used as control variables in analyses: age (in years), gender (l=female, 0=male), and socioeconomic status (a standardized index of educational attainment and monthly household income).

Statistical Procedures

First, bivariate group differences between African-American patients with and without diabetes were assessed using t and %2 tests. Second, African-American patients with and without diabetes were compared on their level of depressive symptoms (using the mean CES-D score for severity of symptomology, the percentage with CES-D >16 for presence of elevated depressive symptoms, and the percentage with PRIME-MD major depressive symptoms) using t and X2 tests. Third, we estimated the effects of diabetes, major depressive symptoms, and the combination of diabetes and depressive symptoms (using a multiplicative interaction term) on the utilization of three types of healthcare services (physician visits, ER visits, and inpatient days). These analyses were estimated using Poisson regression for count variables while controlling for demographic characteristics and comorbid chronic conditions.

Category: Diabetes / Tags: African Americans, Depression, Diabetes, primary care, service utilization

Leave a Reply

Your email address will not be published. Required fields are marked *

CAPTCHA image
*