Patient perceptions of healthcare quality and outcomes

Coleman et. al., (2009) Patient Perceptions of Obstetrician-gynecologists' Practices Related to HIV Testing. Maternal Child Health Journal 13: 355-363.

What were the objectives and hypothesis of the study?

The Coleman and colleagues (2009) study had four main objectives. The researchers sought to determine (1) the number of patients who were tested for HIV by their obstetrician-gynecologist (OBGYN); (2) investigate the attitudes of knowledge of patients concerning HIV testing and their personal risk status; (3) elucidate the key reasons that lead to declining an HIV test; and (4) examine how patients recall their OBGYN's approach to HIV testing.

The researchers hypothesized that women in their sample who were pregnant, seeking preconception care or who were women with risk factors for HIV infection would demonstrate recall for their OBGYN recommending an HIV test.

What was the overall goal/recommendation of the study?

The studying was predominantly seeking to determine whether or not female patients have good recall for remembering the recommendations made by their OBGYN concerning HIV testing. The study hypothesized that women visiting their OBGYN would remember the information provided by their OBGYN in the process of recommending HIV testing, the but results of the study indicated that that majority of women did not recall this process. The researchers suggest that this lack of recall is indicative of a need to re-examine the process through which OBGYN's approach the topic of HIV testing with their patients (Coleman et al., 2009).

3. State the specific study design that was implemented in this study?

The study design used in this study was a survey. Participants, as well as their OBGYN's were provided with surveys to complete. The sampling procedure was not random (Coleman, 2009).

4. Describe what is a continuous variable and list at least 2 continuous variable used in the article.

A continuous variable is one in which the response can have any possible value (i.e. It is not limited to only whole numbers) (Norusis, 2008). One continuous variable used in this study was age, which can include responses that are not limited to whole numbers. The only other continuous variable addressed in this study was household income, which theoretically can be considered a continuous variable if a participant is allowed to provide an actual response (as opposed to selecting a range of responses, which would make it an ordinal variable). It appears, however, that the authors are treating "pregnancy status" as a continuous variable, as they have tested it using t-tests. In order for this variable to be considered a continuous variable, pregnancy status must refer to more than simply a dichotomous variable of being pregnant, or not pregnant.

5. List all the statistical test that were used in this study, giving a brief definition of each and classify each as either a multivariate or bivariate statistic.

Independent Samples T-Test

An Independent Samples T-Test can be used to compare the means of two separate samples. This test is considered a bivariate test (Urdan, 2005).

Principal Component Analysis

PCA is a form of factor analysis that reduces the number of variables in a data set. It works by creating factors, which are groupings of variables that are correlated with one another. The individual factors then do not correlate with each other. The original variables in the data set will have eigenvalues which indicate the extent to which they correlate with each factor. In general, if a variable has an eigenvalue of 1 or higher, it is usually considered to be a part of that particular factor. This is considered a multivariate procedure (Urdan, 2005).

Linear Regression Analysis

Linear Regression allows researchers to predict one variable (Y) from the value of a separate variable (X), or in some cases, multiple variables. Researchers in this study used both age and pregnancy status to be predictor values of HIV testing. The results of a regression analysis will produce a linear equation that can be used to predict values of Y. The regression conducted in this study was bivariate, or simple regression, as they conducted separate regression analyses for each predictor variable. It is possible, however, to conduct a multiple regression analysis, which would then be a multivariate analysis (Urdan, 2005).

Chi Square

A Chi-Square test is used to test the expected values against the actual values measured in a particular sample. The test compares the observed distribution of frequencies to an expected distribution of frequencies, and then determines whether or not these distributions are similar (and likely to be from the same population). In this sense, it is similar to a t-test (Urdan, 2005).

6. What type of bias was evident in this study?

Sampling bias was a significant factor in this study. The study represents the experiences of patients of only a selection of OBGYNs, and the subset of respondents from each OBGYN may not be representative of that particular OBGYN's practice, as the selection process was not random. Because there was not a process for random selection of physicians or of patients, the results of the study may not be generalizable to the population as a whole. The physicians selected were members of an organization that promotes participation in research and clinical studies, and thus these physicians may differ in some unknown, yet systematic way from the entire population of OBGYN's in the country (Coleman, 2009).

7. Summarize the findings displayed in Table 4.

The findings displayed in Table 4 emphasize the difference between patient recall of HIV testing recommendations made by their OBGYN and the reported recommendations by the physicians themselves. The table indicates that a much higher number of physicians reported recommending an HIV test than what was indicated by the actual patients, thus indicating that the patients do not have good recall for the recommendation process. The table also breaks up the responses between pregnant and non-pregnant patients (Coleman, 2009).

8. In the Results section of the paper, under the Factors Influencing HIV Test Recommendations subheading, the author notes that a Chi-Square test revealed that race was significantly associated with patient recall of their Ob-Gyn recommending HIV testing. Explain the equation X2 (3) = 17.8, P < 0.01, in other words:

• What does X2 (3) mean? -- This means that the degrees of freedom is equal to 3.

• What does the 17.3 value indicate? -- 17.8 is the Chi-Square value.

• What is the significance of P < 0.01? This lets us know that the difference in proportions found by the Chi-Square test is significant, because the p value is less than .01. This means that there is a less than 1% chance that these results are due to chance, as opposed to being due to an actual difference between the expected and observed frequencies.