Page 25 - CleanScience_Fall22
P. 25
About the Authors
Carl Grimes,
David Kattari
SYNOPSIS
Carl Grimes
David Kattari
A free website questionnaire was developed for the purpose of exploring and cataloging selected characteristics of resi- dential structures, occupant behaviors, and self-reported symptoms. The exploratory nature of the project was inten- tionally open-ended in order to provide a more complete picture of the three distinct areas of exploration.
With over 300 variables and 80,000 responses, analysis and synthesis of the data is a daunting task. Statistical analysis, however, can help point to certain possibilities for further exploration. This paper focuses on data that indicates the possible relationship between dust and health symptoms.
Dust was not an area of particular focus in the development of the questionnaire. Only two questions were about dust. One question asked about dust on interior surfaces and the other question asked specifically about dust on windowsills.
As analysis of the overall data developed, associations with the reported presence of dust began to appear. This unanticipated emergence of a relatively minor question took on new meaning when considered along with medical findings about dust as a significant factor in human health.
Implications for health effects from dust inhalation have taken on increased concern with findings about the penetration of the ultra-fine fraction of the PM2.5 particles through the lung tissues directly into the bloodstream. The particles are then distributed throughout the body, potentially affecting internal organs including the brain.
Three methods of logistic regressions were run on data involving dust to explore potential associations. The information was “crowd sourced,” so the reported data are subjective. However, subjective information can still be useful if managed transparently, especially given the large size of the data set. Following are notable results.
Symptom Type Logistic Regressions: Surface dust is a statistically significant contributor to all 23 symptoms, the second most impactful condition, and it is a particularly strong contributor to fatigue, sinus, aching, mood changes, and coughing.
Number of Symptoms Linear Regression: Surface dust was the #1 predicter of number of symptoms. Respondents who indicated surface dust reported 0.506 more symptoms than those who did not, and “notice odors” and “sill dust” were the #2 and #3 predictors of number of symptoms.
Dust Logistic Regression: A Logistic Regression examined what factors contributed to dust. Two factors influenced the likelihood of a respondent indicating surface dust: three or more pets resulted in 24% greater likelihood, and a tenure of less than one year in the house resulted in less likelihood.
Next Steps
All three methods of analysis of dust significantly point to some type of involvement of dust with reported symptoms in homes. An investigation and analysis appropriately struc- tured for traditional statistical analysis is needed to advance the information from this exploratory questionnaire.
Perhaps a working hypothesis that a multidisciplinary study of housecleaning procedures, exposure measurements, and medical monitoring would be expected to show a decrease in symptomology from a reduction in dust exposure. If the hypothesis is supported with evidence from studies, the cleaning industry would have a strong, even compelling, opportunity to develop procedures and standards for highly effective dust management, removal, and prevention of accumulation in homes.
FALL 2022
THE JOURNAL OF CLEANING SCIENCE | 25
