The Legacy of Tuskegee and Trust in Medical Care: Measures

Knowledge of the Tuskegee study was assessed by six items. The first five items were multiple-choice, with one correct response for each item. These survey items assessed factual information about the Tuskegee study. The final item asked if a similar study was possible today.

Medical mistrust was assessed using the seven-item Medical Mistrust Index (MMI). The scale employs Likert-type response codes ranging from “strongly disagree” to “strongly agree.” Examples of items included in the mistrust scale are: “Patients have sometimes been deceived or misled by healthcare organizations” and “Healthcare organizations put the patient’s health first.” The MMI shows good reliability (Chronbach’s a = 0.76).

Other measures used in the study were race, age, sex, education, income and insurance status. Age, education and income were analyzed as continuous variables. Race and sex were analyzed as binary variables. Insurance status categories were private, Medicaid, Medicare and uninsured. For analysis, three dummy variables were created for insurance status with private insurance as the reference group.

RESULTS

Respondent Characteristics Table 1 shows the demographic profile of the sample, which is not unlike the distribution of demographic characteristics for the city of Baltimore. The sample was 69.1% African-American and 71.3% female. There was a broad age range within the sample, although 51.2% of the sample was between the ages of 25 and 54 years. Income ranged from <$5,000 to >$60,000. Most of the sample (48.3%) reported their income to be <$25,000, although 34.8% reported an income between $25,000 and $59,000. The remaining 16.9% reported an income of >$60,000.

Nearly one-third of the sample (32.7%) reported graduating from high school, 25.6% had less than a high-school education. The rest of the sample reported having graduated from college (22.1%) or having completed some college (19.6%) without graduating. The largest proportion of the sample reported having private health insurance (50.9%), followed by Medicare (32.9%), Medicaid (21.9%) and those reporting no health insurance (23.9%). It should be noted that due to the overlap of the private, Medicare and Medicaid categories among respondents, insurance status proportions sum to more than 100%.

We first asked respondents if they had ever heard of the Tuskegee study. Respondents who indicated they had heard of the study were asked the items that assessed factual information about the study. Respondents who were unfamiliar with the study were read a brief description (taken from the CDC website) and then asked if they thought a study like the Tuskegee study could happen today.

Approximately two-fifths of the total sample (41.9%, n=168) had heard of the Tuskegee study. However, there were no significant race differences in awareness of the study, with similar proportions among black (41.7%, n=70) and white (44.6%, n=75) respondents reporting having heard of the Tuskegee study. Knowledge of specific aspects of the Tuskegee study was further assessed among those who were aware of the study.

Table 2 shows the results of responses to the knowledge questions regarding the Tuskegee study. Most participants correctly reported when the Tuskegee study began. However, a large minority of both blacks (24.4%) and whites (31.7%) believed the study began two decades later, in the 1950s. The remaining respondents believed the study began in the 1890s (blacks: 4.9%, whites: 0%) or in the 1970s (blacks: 7.3%, whites: 2.4%).

A larger proportion of whites (46.3%) than blacks (32.5%) correctly answered when the study ended. However, the majority of both black and white respondents answered incorrectly. Similar but small proportions of blacks (6.0%) and whites (4.9%) reported the study came to an end four decades early, in the 1930s. However, substantial percentages of blacks (47.0%) and whites (41.5%) believed the study ended in the 1950s. A somewhat higher proportion of blacks (14.5%) than whites (7.3%) thought the study concluded in the 1980s.
The largest proportions of both black and white respondents underestimated the size of the Tuskegee study. The most common response for blacks was that 75 men participated in the study. The majority of whites thought only 200 men were in the study. Only 19.5% of whites and 22% of blacks correctly responded that approximately 600 men participated in the study.

The Tuskegee study was conducted by the U.S. Public Health Service. The facilities of the Tuskegee Institute (now Tuskegee University) were used for some aspects of the study. Only 26.8% of whites and 25.3% of blacks knew that the U.S. Public Health Service conducted the study. More than 29% of blacks and nearly 27% of whites thought the Tuskegee Institute conducted the study. Large proportions of blacks (40.5%) and whites (43.9%) believed the U.S. Army was the organization that conducted the Tuskegee study. A small proportion of blacks (5.1%) and whites (2.4%) reported that the study was conducted by Johns Hopkins University.

Most respondents believed the men followed during the Tuskegee study were given syphilis by the study team. The vast majority of blacks (75.3%) and just over one-half of whites (52.8%) believed this, although a higher proportion of whites (47.2%) compared to blacks (24.7%) correctly indicated the men followed during the Tuskegee study “already had it [syphilis].” A substantial percentage of white respondents (47.2%) believed such a study is possible today. However, a significantly higher proportion of black respondents (76.6%) believed a similar study could occur today.

The belief that a similar study could happen again is particularly germane to the issue of race differences in medical care mistrust. In Table 3, we examined the relationship among race, awareness of the Tuskegee study and the belief that a similar study could happen again. After hearing about the Tuskegee study, 63.6% of unaware whites believed a similar study could happen again today. However, a much smaller proportion of whites already aware of the Tuskegee study (37.8%o) believed a similar study could happen again today. By contrast, 76.9% of unaware blacks and 87.1% of blacks already aware of the Tuskegee study believed a similar study could happen again today. These findings indicate that, for whites, being told about the Tuskegee study made a tremendous difference in their belief that a similar study could happen again. However, for blacks, being made aware of the Tuskegee study made little difference in their belief that such a study could happen again.

We also examined the relationship between race and the incidence of the belief that a similar study could happen again. The proportion of initially unaware blacks who believed a similar study could happen again was compared to the proportion for their white counterparts. This resulted in a ratio of 1.21 (p<0.05), indicating blacks were 21% more likely to believe a similar study could happen again. A similar comparison was made among those who were aware of the study. This comparison produced a ratio of 2.30 (p<0.05), indicating that blacks were 130% more likely to believe a similar study could happen again. Overall, blacks were more likely to believe a similar study could happen again regardless of their awareness of the Tuskegee study.

We used Ordinary Least Squares Regression analyses from SPSS 12 to analyze the relationship between race and mistrust of the medical care system using the MMI. We also examined whether knowledge of the Tuskegee study is associated with medical mistrust and whether it attenuates the relationship between race and mistrust. We specified a series of regression models.

In Model 1, we tested for an unadjusted effect of race on mistrust. There was a positive association, indicating that blacks had higher scores on the MMI and therefore higher levels of mistrust (b=0.100, p<0.05).

In Model 2, we added sex, education, age, income and insurance status to test for a change in the race effect. Black race remained a significant predictor of mistrust, controlling for the demographic variables and insurance status (b=0.166, p<0.05).

In Model 3, a binary variable indicating whether the respondent had heard of the Tuskegee study was added. Similar to our findings for the previous model, black race remained an independent predictor of mistrust after adjusting for demographic variables and awareness of the study (b=0.171, p<0.05).

Finally, in Model 4, we computed a Tuskegee Knowledge Summary Score by summing the correct answers to the five Tuskegee study questions. Adjusting for knowledge of the Tuskegee study resulted in a small reduction in the strength of the relationship between black race and mistrust. However, the relationship remained significant (b=0.164, p<0.05). Overall, the results indicate that black race remained a significant predictor of medical care mistrust controlling for demographic variables, including income and insurance status, as well as awareness and knowledge of the Tuskegee study.