When are patients lost to follow-up in pre-antiretroviral therapy care? a retrospective assessment of patients in an Ethiopian rural hospital

Between 2010 and 2013, a total of 659 patients were enrolled in pre-ART care. Of these,
626 were eligible for this study based on the inclusion criteria.

Socio-demographic characteristics of the patients

The mean age of participants was 30.6 (SD, 8.7) years, 59.4% were females, and 4.4%
of the females were pregnant. Of the total participants, 58.0% were married, 34.4%
had dependent children below 15 years of age, 39.5% had primary level education, and
67.3% were Orthodox Christians. More than half (50.8%) were from rural communities
and 34.9% were of the Amhara ethnicity. The selected socio-demographic characteristics
of the study participants, by pre-ART outcome, are presented in Table 1.

Table 1. Baseline socio-demographic characteristics of adult pre-ART patients enrolled in pre-ART
care at the Sheka Zonal Hospital, Sheka Zone, southwest Ethiopia, 2010–2013, by pre-ART
outcome

Baseline characteristics of the patients

Of the 626 patients enrolled in pre-ART care, 374 (59.7%) were early stage presenters
and 413 (66%) were linked to care with a baseline CD4 count ?350 cells/ml. The baseline
mean CD4 cell counts were 469.9 (SD, 323.2) cells/mm³ and 879.7 (SD, 349.5) cells/mm³for retained and LTFU patients, respectively.

Of the 79.2% of patients who had experienced at least one type of opportunistic infection
during presentation to care, 341 (54.5%) had been treated with CPT.

Of the total participants, 227 (37.2%) were underweight (BMI ?18.5), 377 (61.8%) were
of normal weight (BMI?=?18.5–24.9), and 6 (1.0%) were overweight (BMI ?25). Regarding
tuberculosis (TB) screening, 625 (99.8%) were screened for TB, of which 106 (17%)
were positive.

In terms of the baseline functional status of the cohort during presentation to care,
578 (92.6%) were at working conditions and the rest (4%) and (3.4%) were at ambulatory
and bedridden conditions, respectively.

Service availability, mode of referral, and disclosure of HIV status

Of the total participants, 424 (67.9%) were in the catchment area of the zonal hospital
and 528 (85.2%) were linked to pre-ART care using the internal mode referral. Regarding
disclosure of HIV status, 368 (59%) disclosed their HIV status, of which 332 (90.2%)
did so to their family members. The rest of the patients (41%) had not disclosed the
status to anyone.

Outcomes during follow-up period

Patients were followed up with pre-ART care from 2010 to 2013 for 319.92 person-years
of observation (PYOs). At the end of the observation period, of the patients who were
enrolled in pre-ART care, 178 (28.4%) were lost to follow-up, 20 (3.2%) were transferred
to another facility, 268 (42.8%) had initiated treatment, 33 (5.3%) had died, and
127 (20.3%) were found to be active in pre-ART care. The overall LTFU rate was 55.8
per 100 PYOs. Among those patients lost to follow-up, 111 (62.4%) were female. The
number of LTFU cases among pregnant women in the pre-ART patient cohort was not reported
during the study period.

Survival analysis of time until LTFU

A total of 626 patients were followed for a median of 6.13 (IQR, 4.57–13.23) months
in pre-ART outcomes. The minimum follow-up time was three days and the maximum was
1,080 days (around 36 months). After a six-month period, 166 (93.26%) patients were
lost to follow-up and this number grew to 171 (96.07%) at the end of one year, starting
from the period of enrolment (see Table 2).

Table 2. Estimated number of LTFU cases among pre-ART patients at the end of observation periods
spanning 3, 6, 12, 24, and 36 months at Sheka Zonal Hospital, southwest Ethiopia,
between 2010 and 2013

The mean estimate of LTFU time for patients who started CPT at pre-ART care during
enrolment was 9.88 (95% CI, 9.5–8.3) months, whereas for patients who were not started
on CPT, it was 5.55 (95% CI, 4.9–6.2) months. The difference in time between the two
categories of patients was statistically significant (p 0.001). Similarly, the mean estimate of time until LTFU for patients who disclosed
their HIV status to their family members was 9.8 (95% CI, 9.4–10.3) months, whereas
for patients who did not disclose their HIV status to anyone when enrolling into care
was 5.0 (95% CI, 4.3–5.7) months. The observed difference was statistically significant
(p 0.001).

At enrolment, the cumulative proportions of patients retained were higher among patients
who started CPT, disclosed their HIV status to their family members, and who had a
baseline CD4???350 (see Table 3).

Table 3. Comparisons of LTFU experiences among pre-ART patients, by CPT status and HIV status
disclosure at enrolment, Kaplan-Meier survival table

Factors associated with time to LTFU

During the bivariate analysis of factors associated with time to LTFU, six explanatory
variables with p-values 0.25 were selected for the multivariate Cox regression model. These included
provision of CPT at enrolment, service availability, CD4 cell count, opportunistic
infections, WHO clinical stage, and HIV status disclosure. Three predictor variables,
CPT at enrolment, baseline CD4 count, and HIV status disclosure, were found to be
independent predictors of time until LTFU (see Table 4).

Table 4. Cox regression analysis for independent predictors of timing until LTFU at the Sheka
Zonal Hospital, southwest Ethiopia, from 2010 to 2013