Abstract:
The research work addresses three objectives. First, use Logistic regression model to predict
unit non response in hospital survey, second, the relationship between item and unit
non-response in Hospital survey. Thirdly, determine the proportion of unit and items
nonresponse in hospital survey. Binary logistic regression model was used to model the
outcome of item and unit non-response with respect to demographic variables such as age,
gender, and marital status, and educational level, occupation status while diagnosis and
operation are sensitive variables. A set of 662 patients records was extracted at Ekiti State
University Teaching Hospital, Ado – Ekiti (EKSUTH), 241 were inpatients represent 36.4%
of the distribution while 421 were outpatients which also represent 63.6% of the total
observation. Also, the research work was investigated and it validated the claims from
literature review. It was discovered that some of the independent variable items has
statistically significant effect on the model why some does not have significant effect on it.
(R2=0.738) indicating a moderate strong relationship of 73.8% between the unit and item
nonresponse. Comparison of Results of Logistic Regression Analyses model using the Unit
nonresponse (outpatients and inpatients) as dependent variable and age, gender, education etc.
as explanatory variables was fitted to set of 662 observations.
