Predictors of health decline in older adults with pneumonia: findings from the Community Acquired Pneumonia Impact Study

Table 2 Final logistic regression model of the most important predictors of self-reported decline in health (n = 187¹)

Predictors of Health Decline	Adjusted Odds Ratio²	95% Confidence Interval
Reported the symptom "no energy"
No	1.00	-
Yes	5.06	(1.93, 12.70)
Before CAP, how much difficulty doing usual daily activities because of physical health³
Quite a lot/couldn't do	1.00	-
Not at all/very little/somewhat	7.62	(2.09, 27.81)
People are allowed to smoke in the home
No	1.00	-
Yes	4.70	(1.31, 16.86)
Has brothers or sisters
No	1.00	-
Yes	3.30	(1.24, 8.81)
Before CAP, how active
Somewhat/not very/not at all	1.00	-
Very active	3.04	(1.15, 8.06)
Reported the symptom "sweats"
No	1.00	-
Yes	2.56	(1.00, 6.64)

Final Logistic Regression Model Statistics:
Rho-square = 0.27 (pseudo R², values between 0.2 and 0.4 suggest a very good fit)
Cox & Snell R-square = .209; Nagelkerke R-square = .357 (i.e. between 20.9% and 35.7% of variance is explained by this model)
85.6% correctly classified
¹Five of the 195 (2.6%) older adults had missing values for one or more of the variables included in the final model.
²Odds ratios for categorical variables represent comparisons with the referent group (OR = 1.00) after adjustment for all other variables in the model.
³SF-8 questionnaire item.

ISSN: 1471-2318