Seven More Patients in Every Hundred: Reporting a Text Reminder Trial's Results the Way Clinic Leaders Read Them
Student Name
American College of Education
RES5303: Research Methods and Applied Statistics in Healthcare
Module 6 Assignment
Instructor Name
October 11, 2027
The Study in Brief
Earlier in this course, the course built a research question about whether reminder texts help new patients reach their first visit at a community behavioral health intake clinic. The clinic, invented for this assignment, then tested the question. Over twelve months, 1,800 adults with a mobile number on file who were scheduled for a first appointment were randomly assigned to receive either a one-way text reminder two days before the appointment or a two-way reminder at the same time that invited them to reply to confirm, reschedule or cancel. Attendance was recorded from the scheduling system, with cancellations more than 24 hours in advance excluded from the denominator.
Of the 900 patients who received two-way reminders, 666 attended their first appointment, 74.0 percent. Of the 900 who received one-way reminders, 603 attended, 67.0 percent. The two-way reminders also freed appointment slots: 108 patients replied to cancel or reschedule, and 71 of those slots were filled by other patients from the waiting list. The numbers are the same in every version of this report; what changes is whether the people who must act on them can understand what they mean.
The Conventional Report
A conventional statistical summary would read as follows. Attendance was higher in the two-way reminder group than in the one-way reminder group (74.0 percent versus 67.0 percent; difference 7.0 percentage points, 95 percent confidence interval 2.8 to 11.2; chi-square p = .001). The relative increase in attendance was 10.4 percent. The number needed to treat was 15.
Every element of that summary is correct, and it is appropriate for a journal or a statistical appendix. It is poorly suited to the clinic's leadership team, which includes a clinical director, a practice manager, a finance officer and two peer support specialists. Several of its terms, confidence interval, p value and number needed to treat, are understood differently or not at all by many readers, and the relative increase of 10.4 percent sounds larger or smaller depending on what the reader assumes it is 10.4 percent of.
Why the Format Matters
Research on how health professionals and patients understand statistics explains the problem. Gigerenzer et al. (2007) documented widespread difficulty among physicians and patients in interpreting health statistics, and they showed that some formats are much easier to understand than others. Relative risk reductions tend to make effects look larger than absolute risk reductions, and natural frequencies, statements such as 7 out of 100, are understood far more accurately than conditional probabilities or percentages of percentages. They argued that transparent reporting, using absolute risks and natural frequencies, is a responsibility of those who present evidence.
Measures such as the number needed to treat were designed to make results more intuitive. Akobeng (2005) explains that the number needed to treat is the inverse of the absolute risk reduction and represents the number of people who must receive an intervention for one additional person to benefit. Even so, the phrase treat is odd when the intervention is a text message, and the concept is clearer when stated concretely: about 15 patients need a two-way reminder rather than a one-way reminder for one more patient to attend. The best format for a result is the one that lets the reader picture the patients behind the number.
The Report for Clinic Leaders
What we did. For one year, we randomly sent new patients either our standard text reminder or a new reminder that let them reply to confirm, reschedule or cancel. Random assignment means the two groups were alike in every way except the reminder, so differences in attendance can be credited to the reminder.
What happened. For every 100 new patients who got the reply-enabled reminder, 74 came to their first appointment. For every 100 who got the standard reminder, 67 came. That is 7 more patients seen for every 100 reminded. Across the 150 first appointments we schedule each month, switching every patient to the reply-enabled reminder would mean about 10 more people starting care each month who would otherwise have been lost. In addition, about 12 patients in every 100 used the reply to cancel or reschedule, and about two-thirds of those slots were given to someone else on the waiting list.
How sure we are. The true difference is very unlikely to be zero. Allowing for chance, it could be as small as about 3 more patients per 100 or as large as about 11 more per 100; our best estimate is 7. Even the smallest likely effect would add several new patients a month.
What it means for the clinic. The reply-enabled reminder costs the same as the standard one on our current messaging contract. It increases the number of people who begin behavioral health care, reduces empty appointment slots and gives the front desk earlier notice of cancellations. We recommend switching all first-appointment reminders to the reply-enabled version next month.
Choosing Visuals for the Same Audience
The written report will be accompanied by one chart, and the choice of chart follows the same principle as the choice of words. A bar chart with two bars showing 74 and 67 percent attended, drawn on an axis that starts at zero, shows the difference honestly. A chart with the axis starting at 60 percent would make the same seven-point difference look like a doubling, which is a common and misleading way to make results look more impressive than they are. A second option, an icon array showing 100 small figures for each group with the attenders shaded, conveys the natural frequency directly and has been recommended for communicating risk to general audiences because it lets readers count rather than calculate.
For this leadership team, the icon array will lead and the bar chart will appear in the appendix alongside the conventional statistical summary. The appendix ensures that anyone who wants the technical detail, such as the finance officer checking the estimate of additional monthly visits, can find it, while the main page stays readable in the few minutes a leadership meeting allows for each item. A result is only as honest as the picture drawn of it, because many readers will remember the picture and not the numbers.
What the Results Do Not Show
A report for leaders should also say what the study cannot tell them. It included only patients with a mobile number on file; patients without one, who may face more barriers to attending, were not studied and may not benefit in the same way. It measured attendance at the first appointment only, not whether patients stayed in care or improved. And the result applies to this clinic's population and messaging system; a clinic with different patients or a different reminder schedule might see a larger or smaller effect. The earlier meta-analysis of text reminders found consistent benefits across settings (Guy et al., 2012), which makes it likely that the direction of the result would hold elsewhere, but not its exact size.
Conclusion
The same result can be reported as a difference of 7.0 percentage points with a confidence interval and a p value, or as 7 more patients starting care for every 100 reminded, about 10 more a month, with a likely range from 3 to 11 per 100. The first version is correct; the second is correct and usable by the people who decide. Reporting results for a healthcare audience means keeping every number accurate while choosing formats that research shows people understand: absolute differences, natural frequencies, uncertainty in plain words and a clear statement of what the result does and does not show.
References
Akobeng, A. K. (2005). Understanding measures of treatment effect in clinical trials. Archives of Disease in Childhood, 90(1), 54-56. https://doi.org/10.1136/adc.2004.052233
Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007). Helping doctors and patients make sense of health statistics. Psychological Science in the Public Interest, 8(2), 53-96. https://doi.org/10.1111/j.1539-6053.2008.00033.x
Guy, R., Hocking, J., Wand, H., Stott, S., Ali, H., & Kaldor, J. (2012). How effective are short message service reminders at increasing clinic attendance? A meta-analysis and systematic review. Health Services Research, 47(2), 614-632. https://doi.org/10.1111/j.1475-6773.2011.01342.x
How this RES 5303 Module 6 example is structured
RES 5303 Module 6 usually asks for results reported the way a healthcare audience reads them; your classroom's instructions decide the audience and format. This example shows the conventional statistical summary first, then explains why that format misleads nonspecialists, drawing on research about how health professionals understand statistics. The rewritten report follows, organized as a leader would read it: what was done, what happened in natural frequencies, how certain the result is, what it means for the clinic and what it does not show. Placing the two versions side by side makes the principle of the module visible.
RES5303 Module 6 questions, answered
What does RES5303 Module 6 usually ask for?
RES5303 Module 6 usually asks students to report statistical results in a form a healthcare audience can use, such as a brief for leaders, a board summary or a poster. Many sections expect results to be translated from technical statistics into plain language without losing accuracy. Your classroom's instructions decide the audience and format.
Should I report absolute or relative differences?
Report the absolute difference first, because it shows how many people are actually affected, and include the relative difference only with its baseline. A relative increase can make a small effect sound large or a large effect sound small, depending on what the reader assumes it is compared with.
How do I explain a confidence interval to non-statisticians?
State the best estimate and the range of values the data make plausible, in the same units as the result. For example, the best estimate is 7 more patients per 100, and the true effect is likely between about 3 and 11 more per 100. Then say what even the smallest likely effect would mean in practice.
Write yours, or have the desk draft it
This paper is an original model document written by our desk, not a submitted student paper and not an official American College of Education document. Read it for the moves, then write your own to the instructions in your classroom. If you want one built to your exact prompt and rubric, the first custom sample is free and arrives in 24 to 48 hours.