Factors Associated with Sputum Ordering in Patients with a positive TB Screen Test
In the bivariate analysis, only patients’ general appearance on the last visit, documented as abnormal (X2 = 9.45, p = 0.002), having more than one (1) TB symptom (X2 = 8.17, p = 0.004) and presence of a presumptive alternative diagnosis (X2 = 8.99, p = 0.003), were significantly associated with sputum ordering for patients with a positive TB screen test (Table 2).
Univariable logistic regression analysis results
Before adjustment, patients with an abnormal appearance (ORu, 4.40; 95% CI, 1.65–11.72), and those with more than one symptom (ORu, 4.22; 95% CI, 1.51–11.76), were significantly associated with higher odds of sputum order. Patients with abnormal appearance or having more than one symptom were over 4 times more likely to have a sputum order than those whose general appearance was normal and those having only one TB symptom. However, sputum order reduced by 73% (ORu, 0.27; 95% CI, 0.11–0.65), for patients with a documented alternative diagnosis when compared to those in whom no alternative diagnosis was documented (Table 2).
Multivariable logistic regression analysis results
After adjustment, the odds of sputum ordering in patients whose general appearance was noted to be abnormal (ORa=3.05; 95% CI, 1.07–1.67, p=0.036) were 3.05 times the odds of sputum ordering in patients whose condition was described as normal. Similarly, patients with more than one symptom were over three times more likely to have a sputum order, compared to those with one symptom (ORa=3.42; 95% CI, 1.15–10.21, p=0.028) were significantly associated with sputum ordering for patients with positive TB screening test. The odds of sputum ordering in patients who had an alternative presumptive diagnosis (ORa=0.36; 95% CI, 0.13–0.86, p=0.023) were reduced by 66% when compared to those with no such diagnosis and this was significant, (Table 2).
Table 2
Bivariate and Multivariate analysis of factors associated with sputum ordering, N=92
|
Bivariate Analysis
|
|
|
Uni-variable logistic regression
|
Multivariable logistic regression
|
Sputum order
|
Factor
|
Present, n(%)
|
Absent, n(%)
|
X2
|
p-value
|
(ORu(95%CI)
|
p-value
|
ORa(95%CI)
|
p-value
|
Age category
|
|
|
|
|
|
|
|
|
<15
|
2(40)
|
3(60)
|
5.29#
|
0.381
|
1
|
|
|
|
15-24
|
0(0)
|
3(100)
|
|
|
0.00
|
0.999
|
|
|
25-34
|
19(63.3)
|
11(36.7)
|
|
|
2.59(0.37-17.98)
|
0.335
|
|
|
35-44
|
12(54.5)
|
10(45.5)
|
|
|
1.80(0.25-12.98)
|
0.560
|
|
|
45-54
|
16(64)
|
9(36)
|
|
|
2.66(0.37-19.10)
|
0.328
|
|
|
≥55
|
4(57.1)
|
3(43.9)
|
|
|
2.00(0.19-20.61)
|
0.560
|
|
|
Sex
|
|
|
|
|
|
|
|
|
Female
|
27(55.1)
|
22(44.9)
|
0.07
|
0.783
|
1
|
|
|
|
Male
|
26(60.5)
|
17(39.5)
|
|
|
1.12(0.49-2.57)
|
0.783
|
|
|
Visit type
|
|
|
|
|
|
|
|
|
New
|
28(56.0)
|
22(44.0)
|
0.12
|
0.733
|
1
|
|
|
|
Continuing
|
25(59.5)
|
17(40.5)
|
|
|
1.16(0.51-2.65)
|
0.733
|
|
|
Symptoms No.
|
|
|
|
|
|
|
|
|
1
|
30(47.6)
|
33(52.4)
|
8.17
|
0.004
|
1
|
|
1
|
|
>1
|
23(79.3)
|
6(20.7)
|
|
|
4.22(1.51-11.76)
|
0.006
|
3.42(1.15-10.21)
|
0.028
|
Appearance
|
|
|
|
|
|
|
|
|
Normal
|
27(45.8)
|
32(54.2)
|
9.45
|
0.002
|
1
|
|
1
|
|
Abnormal
|
26(78.8)
|
7(21.2)
|
|
|
4.40(1.65-11.72)
|
0.003
|
3.05(1.07-8.67)
|
0.036
|
Alt diagnosis
|
|
|
|
|
|
|
|
|
Absent
|
37(71.1)
|
15(28.9)
|
8.99
|
0.003
|
1
|
|
1
|
|
Present
|
16(40)
|
24(60)
|
|
|
0.27(0.11-0.65)
|
0.003
|
0.34(0.13-0.86)
|
0.023
|
#=X2 (fisher’s exact test); Lrtest; Ho: Automatically selected model better fit for the data. X2=0.82 p=0.84. ORu: unadjusted Odds Ratio; ORa: adjusted Odds Ratio. |
Common Alternative Diagnoses and Prevalence of Chest X-ray Uptake
The most common alternative diagnoses documented in the client folders were malaria and respiratory tract infections (RTI). Of the forty (40) patients that had an alternative diagnosis documented, half of them were reported to have RTI. The other diagnoses included meningitis, liver disease and weight loss (figure 3). Additionally, it may be essential to embrace the use of CXR as a screening tool for TB among PLHIV. Investigation of patients with CXR was very low, as only 3 out of 92 (3.3%) of the patients with a positive screen test had a CXR order documented.
Health worker Perceived Barriers to Sputum Ordering and Chest X-ray Use
Four (4) out of the six (6) planned interviews were conducted. The two health workers were not willing to participate. All respondents were female. The health workers were interviewed for their perceptions about the potential barriers hindering screening of HIV infected patients for TB and further testing of those with a positive test. Some of the barriers are described below.
High Patient numbers
High patient numbers resulting into a high workload was one of the barriers identified by the health workers as indicated in the statements below:
“The number of patients in this clinic is high. This sometimes hinders provision of quality services including TB screening and evaluation”
Another health work said:
“The clients here are so many. We cannot do much. As you know like elsewhere in Africa, the doctor/nurse to patient ratio is very low. However, we try to make sure that all patients get tested for TB at least twice a year.”
Inability by the Patients to produce sputum
It was also mentioned several times that patients fail to produce sputum when they are asked to.
“Sometimes a patient comes, and you ask them to provide sputum, but he/she says I can’t produce sputum, I have failed. So, in that case no test will be done”.
Patients who are well will not be willing to provide sputum.
“Patients who are well do not want to provide sputum. They don’t see the need for doing many tests. We cannot force them. However, those who are sick, try their best to provide the sputum for testing.”
Cost associated with Chest X-ray (CXR)
The chest X-ray is one of the tests recommended by the GHS as part of the work up of PLHIV. However, its use was very low, and the cost associated with the x-ray services was cited as the main barrier.
“Unlike sputum, chest X-ray is not free. Wherever the patients do x-rays they have to pay. This limits the uptake of the service as part of TB work up.” Said one of the health workers.
Related to cost, another health worker had this to say:
“Although the chest X-ray is covered under the National Health Insurance scheme, patients have to pay money for the radiologist report to accompany the x-ray. This also makes the x-rays services unaffordable for most of the patients.”
Lack of integration of TB services into HIV care and treatment services
From the interviews, it was apparent that there is lack of integration of TB services into HIV care and treatment services as represented by the following quotation
“There is a chest clinic that takes care of TB patients where patients with TB are sent for treatment and further evaluation. Even TB medicines for TB/HIV co-infected patients are not provided here, but at the chest clinic.”
Infrequent trainings to improve/build capacity of health workers.
Health workers also noted that the frequency of the trainings is low. This affects quality of care especially when new health workers come or policies change. The following quotations represent the issue of low training frequency
“Once in a blue moon, we have workshops. The DOTS coordinator organizes trainings once in a blue moon”
Affirmed by a quotation from another health worker who seemed to stress the need for the trainings with an additional reason being the need to fill knowledge gap which is noted to be limited particularly among students.
“Training are few here. We have not had any in the last three months. They should increase the number of trainings. Aside we the staff, we sometimes have students here whose knowledge may be limited”