• To collect information from consumers regarding their food handling and refrigeration knowledge and practices
• To visually assess the contents and cleanliness of home refrigerators
• To evaluate microbial contamination on surfaces within consumers' refrigerators by use of the microbial ATP (mATP) bioluminescence assay.

• Included in a follow-up of a previous consumer study on food handling and refrigeration knowledge and practices study
• Living in Florida or Tennessee.

• Not in previous consumer study on food handling and refrigeration knowledge and practices study
• Not living in Florida or Tennessee.

Recruitment

Researchers used an earlier comprehensive consumer study on food handling and refrigeration knowledge and practices study to draw participants for a follow up in-home study of subjects living in Florida or Tennessee.

Design

• Participants completed a home refrigeration practices survey
• Conditions of the participants' refrigerators were evaluated by a trained observer
• Cleanliness, fullness and organization of five areas (door, upper, middle shelves, bottom shelves and vegetable bins) of each refrigerator were recorded on a four-point scale and potentially unsafe circumstances were noted
• Several 100cm² areas of each refrigerator (usually meat area (either a compartment or location where meat was stored), bottom shelf and vegetable bin) were swabbed with sterile buffer
• An mATP bioluminescence assay was performed on the swabs to assess microbial contamination.

Dietary Intake/Dietary Assessment Methodology

Not applicable.

Blinding Used

Not applicable.

Intervention

Not applicable.

Statistical Analysis

• Survey responses and checklists results were numerically coded if possible and entered into SPAA-PC
• For certain analyses, RLU Results, reflecting mATP, were recoded into five categories (equivalence to CFU per 100cm²):
  • Non-detectable (less than 10³)
  • Up to 2,000 (10³ to 10⁵)
  • 2,001 to 20,000 (10⁵ to 10⁶⁾
  • 20,001 to 200,000 (10⁶ to 10⁷)
  • Over 200,000 (greater than 10⁷)
• Also used to evaluate the data:
  • Frequency analysis
  • Pearson correlation
  • Chi-Square
  • One-way Analysis of Variance
  • Tukey's Multiple Comparison tests.

Timing of Measurements

• A minimum of two surfaces were swabbed in each refrigerator
• Timing of measurements in all refrigerators not specified.

Dependent Variables

Microbial ATP levels in sampled areas (measured via bioluminesence assay performed by use of a microluminometer NHD Model 3560) and PROFILE®-1 Reagent Kit (New Horizons Diagnostic).

Independent Variables

• Self-reported refrigerator practices including:
  • Handling of cold foods
  • Cleaning frequency
• Recorded condition of consumer's refrigerator with respect to cleanliness, fullness and organization (based on scoring by trained observer using a checklist)
  • In the case of cleanliness, observers used scoring system in which 1=very clean and 4=dirty
• Also recorded circumstances that might allow for cross-contamination of foods, presence of moldy or spoiled food or unsealed containers and other potentially unsafe or unusual conditions within the refrigerator.

Control Variables

None.

• Initial N: 147 subjects and household refrigerators in Florida or Tennessee
• Attrition (final N):
  • 147 subjects (84% female, 16% male);
  • 147 household refrigerators (minimum of two surfaces swabbed in each refrigerator; total number of samples=369)
• Age: For members of households: 31% contained at least one elderly individual; 36% had children; 10% had a toddler or infant
• Ethnicity:
  • 53% White, non-Hispanic
  • 31% African American
  • 14% Hispanic
• Other relevant demographics:
  • 92% of participants had high school diplomas or degrees
  • 84% had a household income of more than $15,000
  • 12% of households consisted of five or more persons
• Anthropometrics: Not applicable
• Location: Florida or Tennessee, US.

Key Findings

• Refrigerator observations:
  • About 78% of refrigerator areas were scored as either very clean or clean, 20% were judged slightly dirty and only 2% were considered dirty
  • Refrigerator doors were judged slightly cleaner than the bottom shelves and vegetable bins
  • Cleanliness scores for each of the five refrigerator areas were correlated with one another in all cases.
• Microbial ATP:
  • Although the highest RLU (relative luminescence units) were observed in the meat storage area and the vegetable bin, variation within each refrigerator location was large
  • Overall, 72% of swabs had detectable mATP, suggesting that the majority of home refrigerators contain viable microbial populations
  • Vegetable bins had fewer non-detectable (14%) and more elevated mATP outcomes (greater than 15%) than other areas swabbed
  • Although meat areas in a number of refrigerators showed elevated mATP, the greatest percentage of samples with non-detectable levels were also in this area.
• Cleanliness score and mATP
  • Microbial ATP on the bottom shelf correlated with the cleanliness score for that area (r=0.210, P<0.05)
  • Cleanliness scores for all the refrigerator compartments except the vegetable bins were correlated (r=0.167 to 0.236, P<0.05), with mATP found on the bottom shelf suggesting that bacteria settle from upper shelves to bottom shelf.
• Reported cleaning practices and mATP
  • About three-quarters of participants frequently cleaned up spills in their refrigerators
  • Refrigerators of consumers who more often clean spills in their refrigerators had greater mATP values on the bottom shelves (r=0.251, P<0.05)
  • A majority of participants often or occasionally cleaned compartments within their refrigerators, but half rarely or never empty and clean the refrigerator
  • Mean mATP was greater in refrigerators that were emptied and cleaned less frequently
  • Microbial ATP was inexplicably low in the vegetable bins of those who never thoroughly cleaned their refrigerators
  • Data from other compartments failed to show a clear relationship between refrigerator cleaning frequency and mATP
  • Microbial ATP in the vegetable bin was correlated with the cleanliness score for that compartment (r=0.252, P<0.01), but mATP was not related to the self-reported frequency of washing the vegetable bins.

Table: Percent of Sampled Locations in Refrigerators by mATP Concentration

• A majority of swabbed surfaces of consumer refrigerators contain detectable populations of bacteria as assessed by ATP bioluminescence, indicating the presence of viable microbial populations in most home refrigerators
• Refrigerator cleanliness scores and mATP results support the hypothesis that contaminants within a home refrigerator may settle to the bottom shelf
• Vegetable bins of home refrigerators commonly showed the highest mATP levels, perhaps due to the storage practices of consumers
• Visual appraisal is not a reliable method of assessing microbial contamination within a home refrigerator, nor are self-reported cleaning practices of consumers reliable in predicting microbial contamination.

• Subjectivity of trained observers' cleanliness scores (lack of validity of scoring tool)
• The authors noted these limitations:
  • ATP bioluminescence results may be altered by the presence of cleaning agents and chemical sanitizers or disinfectants (and about two-thirds of subjects in this study reported using some type of cleaning compound either often or occasionally within their refrigerators)
  • Speculation that some participants may have cleaned their refrigerators before the researchers arrived, even though they had been asked not to do so (This cleaning was apparent to the researchers in a few instances)
  • Visual assessment of cleanliness may not be a reliable indicator of microbial contamination.