Purpose

• Gauge attitudes about and experiences with yeast infections that might affect a person’s attitude towards diagnosing their infection
• Assess the factors that people consider when deciding if and how to diagnose their yeast infection
• Develop design criteria for our paper-based yeast infection detection system and determine possible points of concern

Institutional Review Board Approval

Our survey was approved by Purdue University’s Institutional Review Board Human Research Protection Program (# 31228808) as a Category 2 Exemption, which covers human subjects studies involving survey procedures.

Survey Design

Feedback for survey design was provided by Dr. Bruce A. Craig from Purdue’s Statistical Consulting Service to ensure survey was not biased and easy to follow.

Survey Distribution

Our survey was distributed with Amazon Mechanical Turk, although the survey itself was hosted on Qualtrics in accordance with the University of California-Berkeley Committee for Protection of Human Subject’s recommedation for using Amazon Mechanical Turk for academic surveys.[1]

Studies have shown that the Amazon Mechanical Turk Demographic is diverse, yet compared to the world population, MTurk workers tends to be

• More heavily concentrated in the United States of America and India MTurk users from the United States, compared to the general population of the United States, tend to be

Still, Amazon Mechanical Turk was chosen as a survey distribution method due to it being a fairly cheap, non-labor intensive way to achieve a large sample size without relying upon a convenience sample.

Several studies have found evidence indicating that Amazon Mechanical Turk survey data can be just as valid as other common methods for collecting academic survey data [3].

Demographics

Our final sample size was 239 respondents after eliminating responses with invalid answers, such as those that indicated only a genital yeast infection but indicated oral symptoms.

• Age ranges from 19 to 81, with a mean age of 38 years, a standard error of 0.83, a median age of 35 years, and an interquartile range of 16 years.

• The vast majority of respondents were from the United States of America (85.8%), but 10 total countries were represented. 38 different states from the USA were represented.

• 66.1% of respondents were male and 33.9% were female. A total of 75.32% of females and 23.46% of males indicated that they have had genital yeast infections, and 16.46% of females and 17.29% of males have had oral thrush. This is in line with previous studies which estimated that 75% of women will get a yeast infection at some point during their lifetime,1 and thus provides evidence for the validity of this study.

Attitudes Towards Yeast Infections

When asked how much they agree with the statement “a yeast infection is a sexually transmitted infection, with a score of “0” meaning strongly disagree and “10” meaning highly agree, respondents indicated a mean agreement value of 3.04 and a median agreement value of 2. The standard error was 0.206 and the interquartile range was 5. Overall, the majority of people do not consider a yeast infection a sexually transmitted infection.

Factors that increase a person’s likelihood of considering a yeast infection a sexually transmitted infection are . . .

Gender (Males: 4.34, Females: 2.37, p-value: <0.0001)

Younger age (-0.052 agreement values/year, p-value 0.0013, although R2 was low at 0.043)

Experience with yeast infections (Non-overlapping connecting letters indicates a pairwise t-test with a p-value <0.05.)

Factors that do not affect a person’s likelihood of considering a yeast infection a sexually transmitted infection are. . .

• Country of residence (ANOVA p-value: 0.0591, indicates no group significantly different than the rest of the groups)
• State of residence in the United States (ANOVA p-value: 0.1170)

Yeast ID Marketability

As our detection system is designed to be an over-the-counter (OTC) diagnostic test, we wanted to assess the marketplace and demand for this style of diagnosis. While diagnosis by a healthcare professional was the most preferred option, with 42% of people selecting it as their first-choice diagnostic method, 35% of people selected over-the-counter tests as their preferred diagnostic method, and 51% of people selected it as their second-choice diagnostic method, making it the overall second-choice diagnostic method. As such, due to the prevalence of yeast infections, there seems to be a market for OTC tests, like the one we are creating. All of the people who ranked “other” highly described their “other” option as alternative, “natural” diagnosis methods and treatments. While the over-the-counter was not the most-preferred diagnostic method, there is still a large enough market to justify a demand for testing products such as our own.

A total of 29 people listed not seeking an official diagnosis, and trends in their responses as to why that is there preferred method were analyzed. Nine major themes were apparent in their answers, and about half of them believed that it was easy to diagnose yeast-infections based on symptoms or by consulting online resources and forums. 14% of people wanted to go right to self-treating with over-the-counter products. 10% of people included in their response either that seeking a diagnosis was embarrassing or that they liked the privacy involved in self-diagnosis.

A total of 57 people selected that they preferred seeking a healthcare professional for their diagnoses, and the reasons for their preference was analyzed. 43% of respondents indicated that they just overall had a high amount of trust for doctors. 32% of respondents included in their response that they wanted a doctor’s recommendations for treatments or prescription-strength treatments. 16% believed that a doctor was most likely to give them the most accurate diagnosis, and 11% thought that a doctor would give them the most direct answer as to what was wrong. 7% believed going to a doctor was the safest option.

A total of 49 people selected that they preferred using over-the-counter diagnostic tests, and the trends for the reason for their preference was analyzed. 33% of respondents listed the convenience and ease-of-use as reasons why they preferred OTC-diagnoses, with 10% mentioning that OTC-tests were quick or that they thought they would have to wait too long to see a doctor. 12% of people just responded generically that they liked how OTC-tests prevented them from needing to see a doctor, with an additional 6% of people who prefer OTC-diagnostic tests specifying that they believe that it is embarrassing to see a doctor for their yeast infection and 14% liking the privacy of OTC tests. An additional 24% of people preferred OTC-tests due their low cost and the high costs associated with doctors visits. 14% indicated that they like using OTC tests when they already suspect the issue is a yeast infection, and just want to confirm this belief.

Overall, the trends in reasons for preferring certain diagnosis options indicates a need for further education on yeast infections and the risks associated with self-diagnosing and abuse of antifungal creams, as well as a need to end the stigma associated with yeast infections. While getting doctors to support our OTC-test and its accuracy would help, we should have a market as long as our product is easy-to-use, cheap, and quick.

Integration into Paper Assay Design

Our survey was designed to establish the following design criteria.

Our test must take less than 30 minutes to achieve a result.

Our test must cost less than $10.

Our test kits should allow for consumers to analyze two individual swabs to ensure their confidence in their diagnosis.

We also asked the survey respondents if they had any concerns with our proposed product, and discovered the following.

• Test accuracy should be greater than 90%.
• With an internal expiration spot on the test strip, there is concern as to buying a test and then discovering it is expired. A solution for this could occur in our packaging, with an amber cellophane window on the packaging allowing customers to view the expiration spot.
• Over time, the negative control will also react and become blue. There is concern about knowing the optimal time to view the test result.
• Interpreting the test requires comparing the degree of blue of the test spot, negative control, and positive control. There is concern about how accurately users would be able to do so with their naked eyes.
• There are four different spots (expiration, test sample, positive control, and negative control) that can turn blue. There is some concern over properly identifying which spot is which.

References

1. University of California - Berkeley. (2018). Mechanical Turk (MTurk) for Online Research.
2. Ross, J., Irani, I., Silberman, M. Six, Zaldivar, A., and Tomlinson, B. (2010). "Who are the Crowdworkers?: Shifting Demographics in Amazon Mechanical Turk". In: CHI EA 2010. (2863-2872)
3. Sheehan, K. (2017). Crowdsourcing research: Data collection with Amazon’s Mechanical Turk. Communication Monographs, 85(1), 140-156. doi: 10.1080/03637751.2017.1342043
4. Achkar, J., & Fries, B. (2010). Candida Infections of the Genitourinary Tract. Clinical Microbiology Reviews, 23(2), 253-273. doi: 10.1128/cmr.00076-09