Table 3.
Logistic regression model: Failure to initiate combination antiretroviral therapy in 90 days among eligible patients, Kwara and Niger states of Nigeria, 2009–2011 (n=1174).
| Odds Ratio (95% CI) | |
|---|---|
| Clinic | |
| GBRH (ref) | 1 |
| Kuta | 5.70 (2.99, 10.89) |
| LGH | 1.45 (0.86, 2.46) |
| SBSH | 2.46 (1.58, 3.85) |
| UMYMH | 0.84 (0.40, 1.76) |
| Age (per 5 yrs) | 1.01 (0.94, 1.09) |
| Sex and pregnancy | |
| Female (ref) | 1 |
| Pregnant Female | 0.65 (0.31, 1.38) |
| Male | 1.32 (0.94, 1.87) |
| Education | |
| None (ref) | 1 |
| Started primary | 1.29 (0.59, 2.81) |
| Completed primary | 1.56 (0.91, 2.68) |
| Secondary | 1.43 (0.88, 2.33) |
| Post secondary | 2.08 (1.03, 4.20) |
| Qur'anic | 1.11 (0.58, 2.12) |
| Marital status | |
| Married (ref) | 1 |
| Divorced | 1.25 (0.43, 3.66) |
| Separated | 0.79 (0.33, 1.88) |
| Single | 1.31 (0.68, 2.54) |
| Widowed | 1.36 (0.78, 2.39) |
| Occupation | |
| Employed (ref) | 1 |
| Other | 0.98 (0.52, 1.82) |
| Student | 1.71 (0.61, 4.81) |
| Unemployed | 1.08 (0.66, 1.76) |
| BMI (per 1 kg/m2) | 0.93 (0.88, 0.98) |
| Functional status | |
| Working (ref) | 1 |
| Ambulatory | 1.92 (1.15, 3.18) |
| Bedridden | 4.17 (1.63, 10.67) |
| CD4+ cell count (cells/μL) | |
| 50 vs. 200 | 1.49 (0.93, 2.39) |
| 350 vs. 200 | 2.10 (1.31, 3.35) |
| 500 vs. 200 | 3.02 (1.69, 5.39) |
| Hemoglobin (per 1 g/dL) | 1.00 (0.91, 1.10) |
| WHO clinical stage | |
| I (ref) | 1 |
| II | 1.05 (0.60, 1.82) |
| III | 1.12 (0.70, 1.81) |
| IV | 1.16 (0.50, 2.71) |
| Month of enrollment | |
| June 2009 (ref) | 1 |
| December 2009 | 1.01 (0.68, 1.48) |
| June 2010 | 1.18 (0.64, 2.19) |
| December 2010 | 2.13 (1.19, 3.81) |
a
There is evidence that the association between the log-odds of delayed HAART initiation is non-linear with CD4+ count (p=0.021) and date of enrollment (p=0.082). The last day of the month (e.g. June 30) is used to summarize odds ratios for month of enrollment.
b
There are 1174 patients included in this model. Missing values of baseline predictors are accounted for using multiple imputation.
GBRH: Gawu Babangida Rural Hospital; LGH: Lafiagi General Hospital; SBSH: Sobi Specialist Hospital; UMYMH: Umaru Yar'adua Memorial Hospital.