HLTH 4205 Week 8 Descriptive Data Summary Example

Describing the Newly Insured: A Data Summary of 212 Adults at Enrollment and the Reasons Behind Each Statistic Chosen

Student Name

Walden University

HLTH 4205: Introduction to Research Methods and Analysis

Instructor Name

Month Day, Year

Describing the Newly Insured: A Data Summary of 212 Adults at Enrollment and the Reasons Behind Each Statistic Chosen

The Data

This summary describes the baseline survey data from the clinic's study of newly insured adults: 212 adults between 19 and 64 who signed up for Medicaid at the clinic's enrollment desk over a seven-month period and agreed to complete the enrollment survey. The data are a composite dataset built for this course, with values shaped to resemble what such a survey would produce. The summary reports five variables: age, sex, emergency department visits in the past twelve months, having a usual source of care and time since the last primary care visit. For each, it reports counts, a measure of center and a measure of spread, and explains why those statistics were chosen. Reporting guidelines for health research recommend exactly this pairing of each statistic with its justification and its denominator (Lang & Altman, 2015).

Summary Table

Table 1 summarizes the baseline data.

Table 1

Baseline Characteristics of Newly Insured Adults, N = 212 (Composite Data)

Age: Why a Mean and Standard Deviation

Age was summarized with a mean and standard deviation because its distribution was roughly symmetric: the mean, 38.4, and the median, 37, were close, and a histogram showed no long tail. When a variable is approximately normal, the mean and standard deviation describe it efficiently, and about two-thirds of values fall within one standard deviation of the mean (Altman & Bland, 1995). The range is added because the eligibility limits, 19 and 64, are meaningful to readers. The standard deviation, not the standard error, is reported, because the purpose is to describe how much ages vary among participants, not how precisely the mean was estimated (Altman & Bland, 2005).

Sex and Usual Source of Care: Why Counts With Percentages and Their Bases

Categorical variables are reported as a count over its base followed by the percentage, such as 86 of 210. Showing the base matters because two participants did not answer the usual source of care question, so the percentage uses 210, not 212. A reader who saw only 41.0% could not tell whether missing answers were counted as no. The percentage is given to one decimal place, which is enough precision for a sample of this size; more decimals would suggest accuracy the data do not have.

Emergency Visits: Why a Median and Interquartile Range Beside the Range

Emergency department visits were heavily skewed: most participants had zero, one or two visits, while a few had many, up to fourteen. The mean was 1.7, higher than the visits of three-quarters of participants, because a handful of frequent users pulled it upward. The median, 1, describes a typical participant better, and the interquartile range, 0 to 2, shows where the middle half of participants fall. The full range is added because the frequent users matter for planning, and a reader should know they exist. Reporting guidelines recommend medians and interquartile ranges for skewed data for this reason (Lang & Altman, 2015).

Time Since Last Primary Care Visit: Why Missing and Never Are Reported Separately

This variable required three decisions. First, 33 participants said they had never had a primary care visit as an adult. They are not missing data; they are an answer, and an important one. They cannot be given a number of months, so they are reported separately rather than dropped or assigned an arbitrary value. Second, 8 participants did not answer, and they are reported as missing. Third, among the 171 who gave a time, the distribution was skewed, with some reporting visits many years ago, so a median and interquartile range are used. Combining never with missing, or excluding the never group silently, would have made the newly insured look more connected to primary care than they are.

What Was Not Summarized and Why

The survey also asked about income, race and ethnicity, language and chronic conditions. Those variables are reported in a separate table for the full study, because this summary was limited to the five variables the six-month follow-up will compare. Two further decisions deserve mention. Income was collected in bands rather than dollars, so it can only be reported as counts in each band, not as a mean or median. And the chronic conditions list allowed more than one answer, so its percentages will add to more than 100, which the full table must say in a note. Leaving these out of a short summary is reasonable; leaving them out without saying so would not be.

What the Summary Shows

At enrollment, the typical participant was in their late thirties, slightly more likely to be female, had visited an emergency department once in the past year and did not have a usual source of care. Fewer than half had a usual source of care, and about one in six had never had an adult primary care visit. These figures describe the 212 participants at one point in time. They do not show whether coverage will change them, which is the purpose of the six-month follow-up, and they cannot be generalized beyond adults who enrolled through this clinic's office and agreed to take part.

Why the Arithmetic Was the Easy Half

Every number in the table took seconds to calculate. The work was in choosing which number to show and saying why. A mean for emergency visits would have been arithmetically correct and descriptively misleading, and a percentage without its base would have hidden two missing answers. The choices, not the calculations, decide whether a reader comes away with an accurate picture of the people in the study.

References

Altman, D. G., & Bland, J. M. (1995). Statistics notes: The normal distribution. BMJ, 310(6975), 298. https://doi.org/10.1136/bmj.310.6975.298

Altman, D. G., & Bland, J. M. (2005). Standard deviations and standard errors. BMJ, 331(7521), 903. https://doi.org/10.1136/bmj.331.7521.903

Lang, T. A., & Altman, D. G. (2015). Basic statistical reporting for articles published in biomedical journals: The "Statistical Analyses and Methods in the Published Literature" or the SAMPL guidelines. International Journal of Nursing Studies, 52(1), 5-9. https://doi.org/10.1016/j.ijnurstu.2014.09.006