5.2 - Summarizing the Data

Data from a survey that includes indicators of the presence or absence of disease and a risk factor can be summarized as shown below:

Table 2: 2 × 2 Table for an Epidemilogic Cross-Sectional Study

  Case (Number) Non-Cases (Number) Total Exposure (Number)
Exposed A B \(\text{Total}_{\text{Exposed}}\)
Not Exposed C D \(\text{Total}_{\text{Not Exposed}}\)
  \(\text{Total}_{\text{Cases}}\) \(\text{Total}_{\text{Non-Cases}}\) Total

Measures of Disease Frequency

  • Disease Prevalence, by Exposure Status
    • For Exposed: A/(A+B)
    • For Not Exposed: C/(C+D)
  • Exposure Prevalence, by Disease Status
    • For Cases: A/(A+C)
    • Fon Non-cases: B/(B+D)
  • Odds of Exposure, by Disease Status
    • For Cases: A/C
    • For Non-cases: B/D

Measures of Association

[A/(A+C)] / [B/(B+D)]

[A/(A+B)] / [C/(C+D)]

[A/(A+C)] - [B/(B+D)]

[A/(A+B)] - [C/(C+D)]

[A/C] / [B/D] = [A*D] / [B*C]

  • Ratio of Exposure Prevalences
  • Ratio of Disease Prevalence
  • Difference in Exposure Prevalence
  • Difference of Disease Prevalence
  • Exposure Odds Ratio

Logistic regresssion would be an appropriate statistical method because the outcome is binary (case, non-case).. The important point is that since this table shows data from a survey, it provides only only a 'snapshot' of the situation. Time-to-event methods of analysis would not be applicable from such data.

Summarizing the data from San Pablo:

In San Pablo, 75 households were entered into the sample. Of these, 62 (82.7%) completed a household survey, 11 were not at home or the house was abandoned. Only 2 households refused to participate in the survey. Thirty-seven (59.7%) of responding households used piped water and 25 (40.3%) used bottled water; 51 (82.3%) of households subsequently boiled water that was to be used for drinking purposes. Water used for cooking and washing purposes was typically not boiled or treated additionally. Limited money was frequently reported as a condition which prevented residents from seeking medical care. Perhaps because of the large percentage of households that boiled drinking water, there was a generally low prevalence of acute and chronic diarrheal disease. Respiratory conditions however, were more prevalent than anticipated.

Might there be other reasons the survey results indicate a lower than expected prevalence of diarrheal disease?