HCS 465 Week 2 Sampling and Data Collection Example

Who Gets Asked and How: Critiquing the Sampling and Data Collection Behind a Health System's Claim That Its Patients Love Online Scheduling

[Student Name]

University of Phoenix

HCS/465: Health Care Research Utilization

Week 2 Assignment

[Instructor Name]

[Date]

The health system, its survey and its figures are composites written for a model paper.

The marketing department of a regional health system recently reported that 91% of patients were satisfied or very satisfied with its new online scheduling tool, and leadership is considering removing most phone scheduling staff. The report's method section is three sentences long: a pop-up survey appeared after patients finished booking an appointment online, 1,240 patients answered over two months, and the question read, "How much did you enjoy our convenient new scheduling tool?" This paper critiques the sampling and data collection behind the claim and proposes a stronger design.

Who Was Sampled

The sampled population was people who successfully booked an appointment online. That excludes everyone the decision will affect most: patients who tried the tool and gave up, patients who never tried it because they lack internet access or comfort with technology, and patients who prefer the phone. A survey that reaches only people who finished the task is like asking diners who stayed for dessert whether they liked the restaurant; the ones who walked out never get the question. This is a convenience sample drawn from the most satisfied part of the population, and it biases satisfaction upward.

How Data Were Collected

The pop-up appeared inside the tool immediately after a successful booking, a moment of relief and accomplishment. Data collected at the point of success, through the same channel being evaluated, overrepresents positive experience. Patients who had trouble would have had to return to the tool to complain, and few do.

How Many Responded

In two months, about 20,600 appointments were booked online, so 1,240 responses represent a response rate of about 6%. People who answer optional pop-ups may differ from those who close them, perhaps being more engaged or more satisfied. Even with a sample of more than a thousand, a response rate this low leaves the results open to nonresponse bias of unknown size.

How the Question Was Worded

The question asked how much patients "enjoyed" a "convenient new" tool. It assumes enjoyment and describes the tool favorably before asking about it, which is a leading question. A neutral item would ask how easy or difficult scheduling was, on a balanced scale.

Probability and Nonprobability Sampling

Sampling methods fall into two families. In probability sampling, each person in the population stands a known, nonzero chance of being picked, as in simple random sampling or stratified random sampling, which divides the population into groups and draws randomly within each. Probability samples support generalizing to the whole population. In nonprobability sampling, selection depends on availability or judgment. Etikan et al. (2016) compared the two most common nonprobability methods: convenience sampling, which takes whoever is easiest to reach, and purposive sampling, which deliberately selects people with particular knowledge or experience. Both are useful, especially in qualitative work, but neither supports claims about how common a view is in the whole population (Creswell & Creswell, 2018).

A Stronger Design

The population of interest is all patients who scheduled a primary care or specialty appointment in the two months, by any channel. The redesign draws a stratified random sample of 1,200 patients, stratified by age group and by scheduling channel (online, phone or in person), so that older patients and phone users are represented in proportion or oversampled for analysis. Each sampled patient receives a short survey by text or email with a mailed paper version and one reminder call, a mixed approach that raises response rates among people who do not answer online surveys. The survey asks how easy or difficult it was to schedule, how long it took and whether the patient needed help, with balanced response options, and it asks patients who did not use the tool why not.

Adding Depth With Interviews

Numbers will show how satisfied different groups are but not why. A purposive sample of 15 to 20 patients who abandoned online scheduling, or who called for help, will be interviewed by phone about what went wrong. This small qualitative component can explain a disappointing number and point to fixes. Interviewees will be chosen to include older patients, patients who prefer Spanish and patients with disabilities, since their experiences are the most likely to be missed by any survey.

Measuring What Matters to the Decision

Satisfaction is only one outcome the decision depends on. Leadership also needs to know how many patients who start online booking finish it, how many then call anyway, how long phone callers wait and whether no-show rates differ by channel. Much of this comes from system logs rather than surveys: the scheduling tool records abandoned sessions, and the phone system records wait times and hang-ups. Combining log data with survey data gives a fuller picture than either alone, and log data have no response rate problem because they capture every attempt.

Practical Limits

The redesign costs more than a pop-up: printing, postage and staff time for calls, perhaps $6,000 in total. A random sample also needs a current list of all patients who scheduled, which the scheduling system can produce in minutes. Compared with the cost of a scheduling decision that affects every patient, the expense is modest, and access to care, including how easily people can get an appointment, is a basic measure of how well a delivery system works (Shi & Singh, 2022).

Questions a Manager Should Ask

Before acting on any survey, a manager should ask who was invited, who answered and who did not, how and when the question was asked, and whether the wording steered the answer. In this case, those questions reveal that the 91% figure describes satisfied online bookers, not patients as a whole.

Conclusion

The health system's satisfaction figure rests on a convenience sample of successful users, collected at the moment of success, with a 6% response rate and a leading question. Each choice pushed results toward satisfaction. A stratified random sample, mixed-mode data collection, neutral wording and a few purposive interviews would tell leadership what it needs to know before reducing phone scheduling.

References

Creswell, J. W., & Creswell, J. D. (2018). Research design: Qualitative, quantitative, and mixed methods approaches (5th ed.). SAGE Publications.

Etikan, I., Musa, S. A., & Alkassim, R. S. (2016). Comparison of convenience sampling and purposive sampling. American Journal of Theoretical and Applied Statistics, 5(1), 1-4. https://doi.org/10.11648/j.ajtas.20160501.11

Shi, L., & Singh, D. A. (2022). Delivering health care in America: A systems approach (8th ed.). Jones & Bartlett Learning.