Journal of Behavioral Health and Psychology. 2014;
3(2):(101-347)
Patient and treatment related variables as predictors of length of stay in a Nigerian Neuropsychiatric Hospital
Adegoke Oloruntoba Adelufosi, Kehinde Fakorede, Ayodeji Obasan, Ibrahim Olajide Opeewe, Tunde Masseyferguson Ojo
Abstract
Aim: The aim of this study was to assess the hospital length of stay (LOS) of patients discharged from the
acute wards of a psychiatric hospital in Nigeria and to determine its relationship with demographic, illnessrelated
and treatment related variables. Methods: This was a retrospective cross-sectional study involving
case note review of patients discharged from the acute psychiatric ward of the hospital during 1-year period.
A questionnaire was used in collecting data about the socio-demographic variables (e.g. age, gender, distance
of place of abode from hospital), clinical characteristics (diagnosis, co-morbid physical illness etc.) and clinical
practice or treatment related factors (e.g. frequency of review by consultant psychiatrists, administration of
electroconvulsive therapy etc.). Results: A total of 93 patients’ case notes were analyzed, comprising 87.7%
of the total number of discharges during the study period. The hospital LOS was 55 days. Using the median
LOS as the cut-off, 49.5% had short LOS (LOS <55 days), while 50.5% had prolonged LOS (LOS ≥55 days).
LOS was significantly associated with employment status (χ2 = 3.871, P = 0.049), distance of patients’
residence from the hospital (χ2 = 4.300, P = 0.038), reported medication side effects while on admission
(χ2 = 3.986, P = 0.046), previous psychiatric hospitalizations (χ2 = 4.748, P = 0.029) and duration of
current illness before admission (Z = −4.020, P < 0.001). The result of the binary logistic regression analysis
showed that only previous psychiatric hospitalizations made a unique statistically significant contribution
to the model (beta = 1.021, wald = 4.433, P = 0.035, odds ratio = 2.775, 95% confidence interval =
1.073-7.174). Conclusion: Despite their significant association with hospital LOS, socio-demographic, illness
and treatment related factors explained only a minimal variation in hospital LOS. It is possible that variables
“external” to either patients’ characteristics or physicians’ practice style, are more significant predictors of
long hospital LOS. These findings may help guide policy makers in improving the efficiency mental health
service delivery in Nigeria.