Diagnostic Delay among New Smear Positive Pulmonary Tuberculosis Patients in Amhara Region, Northwest Ethiopia: A Two-Perspective Analysis Solomon Abebe Yimer Supervisors: Professor Gunnar Bjune MD, PHD CO - Supervisor Associate Professor Getu Degu University of Oslo Faculty of Medicine Department of General Practice and Community Medicine
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Diagnostic Delay among New Smear Positive Pulmonary Tuberculosis Patients in Amhara Region, Northwest Ethiopia: A Two-Perspective Analysis
Solomon Abebe Yimer
Supervisors: Professor Gunnar Bjune MD, PHD
CO - Supervisor Associate Professor Getu Degu
University of Oslo Faculty of Medicine
Department of General Practice and Community Medicine Section for International Health
University of Oslo
June/2004
Thesis Submitted in Partial Fulfillment of the Master of Philosophy Degree in International Community Health
Abstract
Delay in the diagnosis of tuberculosis (TB) causes more severe illness, more
complication and an increased period of infectivity in the community. A study in Amhara
region in 2001 showed that, among those who had history of cough of more than 3
weeks, only 30% visited the formal health care facilities. We hypothesized that there was
a significant patients’ and health systems’ delay in the diagnosis of pulmonary TB in
Amhara region, and this study was conducted to test our hypothesis.
Objectives: To determine and analyze the length and associated risk factors of patients’
health providers’ and health systems’ delay among new smear positive pulmonary TB
patients in Amhara region, Northwest Ethiopia.
Methods: Within the setting of government health care facilities in Amhara region, we
conducted a cross-sectional study from September 1 - December 31/2003. A total of 384
new smear positive pulmonary TB patients participated in the study. Patients were
interviewed on the same date of the diagnosis using a semi-structured questionnaire.
Result: The median total delay was 80 days (IQR 44-130 days) and the median patients’
delay was 30 days (IQR 15-90 days). Forty eight percent of the subjects delayed for more
than one month. The median health providers’ and health systems’ delays were 61 and 21
days, respectively. In logistics regression, home distance >10 Km to a medical provider
(adjusted odds ratio [ORadj] 3.81, 95% confidence interval [CI] 2.21-6.57) and self-
Socio-economic variables: occupation, educational status, income and distance
from health facility
Health service factors, formal health providers, non-formal health providers,
presenting symptoms
2.5.4.1 Operational definitions of variables
TBMU: A health care facility where microscopy for AFB
is done and anti -TB chemotherapy is initiated.
Total delay: The total time (measured in days) from
reported onset of major symptoms (cough,
haemoptysis) to commencement of anti-TB
treatment.
Health systems’ delay: The period (measured in days) from first visit to
a medical provider to first initiation of anti-TB
treatment.
Medical providers’ delay: The period (measured in days) from first visit to
a medical provider to reporting to a TBMU
Health seeking period: The period (measured in days) from onset of
the major symptoms of TB to first visit to a
health provider
Patients’ delay: The period (measured in days) from the onset
of the major symptoms (cough, haemoptysis,
etc) to first visit to a medical provider.
Health provider: Any person consulted by the patient about
his /her sickness that gave or prescribed some
thing (whatever the form) for treatment. This
does not include the family.
Health providers’ delay: The time (measured in days) from the first
contact to a health provider to the first time
anti-TB chemotherapy is initiated
Diagnosing facility’s delay: From the time (measured in days) the patient reported to a diagnosing facility till the first time the patient starts anti-TB treatment.
Formal health providers: Modern health care facilities; such as health
centers clinics, hospitals either government or
privately owned.
Non-formal health providers: These include traditional healers, herbalists
and religious healers, (holy water) and drug
retail outlets.
.
Traditional health providers: These are traditional healers, herbalist,
religious healer (holy water).
Drug retail outlets: These are pharmacies, drug stores, rural drug
venders and open market drug sellers
Income: Income was divided into 4 categories; no
income, irregular income, regular income 1-300
Birr and regular income above 301Birr.
Housewives, students and the unemployed were categorized as no income group. Farmers are categorized in the irregular income group and self-employed were categorized in either of the two regular income group by calculating their total income on monthly basis.
Knowledge: knowledge is information or fact that an
individual is aware of. In this study, it was
defined based on the awareness of the patient
about the symptoms and treatment of TB.
Stigma: Feeling of disapproval that TB patients
experience due to their illness in their day-to-
day life within the community.
2.5.5 Data quality
As described earlier, questionnaire was used for data collection. The original
English version was translated into Amharic for the actual data collection. The
Amharic version was back translated into English to check the consistency of the
translation. In general, Quality of data was assured through the following
methods;
Careful design and translation of the questionnaire
Pre-testing and standardizing of the questionnaire
Proper training of interviewers
Continuous check ups of data collection procedures through intensive
supervision
Patient register cards and TB registration books and laboratory registries
were cross-checked
2.6 Data analysis
After the data collection was completed successfully, the data were entered into
a computer and routine checking and cleaning were preformed. The statistical
package Epidemiology Program Office (Epi-Info) version 6 and statistical
package for the social sciences (SPSS) version 11.0 were used for analysis.
Percentages and proportions were calculated to show the distribution of the
population by socio-demographic characteristics. To determine the lengths of the
patients’, health systems’ and total delay; medians, means, inter-quartile range
were applied.
When assessing the risk factors for patients’, health providers’ health systems’
and total delay, the median delay period was used to dichotomize the sample in
to either shorter or longer delay period. Group differences were compared using
Mann-Whitney, Kruskal-Wallis (more than two groups) and chi- square tests.
Ninety-five percent confidence interval and odds-ratio were used to assess the
associated risk factors of the different delays. A p-value of less than 0.05 was
considered statistically significant. Finally, after all the potential covariates were
first identified by univariant analysis, logistic regression analysis was performed
to assess the relative impact of predictor variables on the outcome variables.
When assessing knowledge of respondents about TB, six questions were posed,
and a score was calculated from the awareness part of the questionnaire.
Awareness regarding TB was assessed from two angles, i.e. their knowledge
about treatment of TB and the seriousness of the disease. In the treatment part,
subjects were asked three questions including whether TB is curable or not, fee
for treatment and duration of treatment. If they believed that TB can be cured, the
assumption is that they might be willing to seek treatment. Concerning the
seriousness of the disease, subjects were asked about causes of TB, risk of
patients and people around them if they were not treated. For each of the six
questions, a value one was given if responded correctly and a value zero if
responded incorrectly. Then, the mean and the inter-quartile scores were
calculated. Finally, the score was divided into high knowledge and low
knowledge. Respondents that fall in to the third quartile were given a value one
and considered to have high knowledge and the other given value two
considered to have low knowledge. Then it was cross-tabulated with the main
outcome variables for possible associations.
When assessing stigma associated with TB/HIV, six questions were again posed
to the respondents. Each question was given a value one if answered correctly
and a value zero if answered incorrectly. Then the mean and inter quartile scores
were calculated. Respondents that fall in to the third quartile were given a value
one and considered to have less stigmatizing experiences and the others given
value two and considered to have high stigmatizing experiences. Then, it was
cross-tabulated with the dependent variables to look for possible association.
2.7 Communication of results
The study result will be presented as a master’s thesis at the institute of general
practice and community medicine at the University of Oslo, Norway. One or two
articles presenting the results will be submitted to international journals and local
journals in Ethiopia, the results of the study will also be presented to the regional
TB control program in the Amhara Regional State Health Bureau.
2.8 Ethical consideration
The project proposal was evaluated by relevant Ethical Committee in Norway
and the Ethiopian Science and Technology Commission. Both bodies have
ethically cleared the project.
Before the interviewing was commenced, the purpose of the study was clearly
explained by the interviewers for every participant of the study including how the
interview was going to take place. Great respect was given to the study subjects.
The patients were fully empowered to decide on their willingness to participate in
the study. Moreover, the study subjects were assured that there would not be
any risk of participating in the study. Their willingness to participate in the study
was confirmed by taking their informed consent. Both written and oral consent
systems were employed depending on the level of education of the study
subjects.
For the purpose of confidentiality, during the analysis of the records, names and
identifying features were coded to protect their privacy and after an interview was
over, the questionnaire was kept in a locked cabinet. The entire interview was
conducted in a private room in the health care facility. Taking into consideration
how interviewing is tiresome for the patient, we were very much careful not to
use more than the intended time of 20-30 minutes. All of the patients recruited
for the study volunteered to participate in our study.
Chapter three: Results
In this chapter, we will start describing the findings of our study first by presenting
the socio-demographic characteristics of the study population and then, we will
mention how the subjects perceived their illness initially and the actions they took
on their own in response to the symptoms. Following this, we will look into the
different delay periods and the associated risk factors. This data is presented in
two ways. In the first part, all health providers are taken as a reference point to
calculate the health-seeking period and the health providers’ delay. In the second
part, only medical providers were considered to calculate the patient’s, medical
providers’, diagnosing facilities’, health systems’ and total delay. We believe that,
this structuring of the result section might give a better understanding of our
material to the reader.
3.1 Socio-demographic characteristicsIn this study, a total of 384 new smear positive pulmonary TB patients were
interviewed from September 1 – December 31/2003. The socio-demographic
characteristics of the respondents are summarized in table 2. The proportion of
males slightly exceeds that of the females with a ratio of 1.19: 1. The mean age
was 29.8 ±10.52 SD years with a median age of 28 years, minimum 16 and
maximum of 70 years. The mean age for males and females was 30.1 and 29.6
years, respectively. Most, 95.9%, of the subjects were in their productive age
(15-54 years old).
Married group constituted the highest proportion compared to the others. There
was sex-related significant difference in marital status among the sample
population (² = (3) 42.2, P< 0.001). Eighty-five out of one hundred eighty two
(46.7%) females were married compared to 55 (27.2%) males. On the other
hand, 98 out of 202 (48.5%) males were single compared to 36 (26.9%) females.
This difference was again statistically significant (² = (2) P= < 0.001). The
majority (97.1%) of the sample population belonged to the Amhara ethnic group
and the proportion of christians was higher compared to the other religions put
together, and the difference was statistically significant (² = (2) P= < 0.001).
Table 2 Socio-demographic characteristic of smear positive pulmonary TB patients in
Amhara region
_____________________________________________________________________Characteristics Number % _____________________________________________________________________Sex
The findings also showed that, fifty six percent of the respondents resided within
10 km radius of a medical facility. As for the type of house used for dwelling,
39.1% lived in a hut and 60.9% lived in an ordinary corrugated sheet iron roof
house. The houses in both cases had an average room number of one. The
average family size was 5, and in 51.3% of the households, the number of
children ranged from 3-5. Among the total respondents, 44.3% did not have any
form of defined income. However, among those who had regular cash income,
the average monthly income was 250 Birr, which is equivalent to USD$ 29.0
based on the current market. In general, the distribution of income among the
sample population showed statistically significant difference (² (3) =146.3 P <
0.01), and specifically when respondents’ income was cross-tabulated with the
sex of the respondents, we found that more females were in the no income group
compared to the males. The proportion was 61% verses 36 % and the difference
was statistically significant (table 4).
Table 4 comparisons of the respondents sex with income among smear positive pulmonary TB cases in Amhara region
_____________________________________________________________________Variable Male female P-Value______________________________________________________________________IncomeNo income 74 111Irregular income 83 31(² (3) =146.3 P < 0.01)Income 1-300 Birr 19 17Income above 301Birr 26 23______________________________________________________________________
In this study, female respondents were asked whether they could decide on their
own regarding where to go for help during their illness, 327 (85%) responded that
they could decide on their own.
Subjects were asked also to describe if they had ever smoked, consumed
alcohol or chewed khat and their response was cross-tabulated against the
duration of illness before seeking help from the medical providers. The result
showed that there was no statistically significant difference among these groups
with regards to patients’ delay (table 5).
Table 5 Chi-square test showing comparisons of patients’ delay with current habits among smear positive pulmonary TB patients’ in Amhara region _______________________________________________________________
Delay No delayCharacteristics ≤ 30 days > 31 days p- value_______________________________________________________________________Smoking
3.2 Initial symptoms, perception of illness and first action
The major symptoms that patients experienced during the course of their illness
are presented in figure 1. As can be seen on the graph, cough was the most
frequent symptom, followed by tiredness. Figure 2 shows chief complaints by
patients. In 60.7% of the cases, cough was also the chief compliant for
presenting to health providers.
86.5 %92.2 %85.2 %90.6 %79.4 %
25.0 %
76.0 %96.4 %
13.5 %7.8 %14.8 %9.4 %20.6 %24.0 %
75.0 %
3.6 %
0%
20%
40%
60%
80%
100%
120%
yes
no
Figure (1): Graph showing the major presenting symptoms among smear positive
pulmonary TB patients in Amhara region.
When subjects were asked to describe what might have caused the symptoms at
the onset of their illness, only 17.5% of the respondents attributed their illness to
TB. Most (82.3%) suspected that they had other diseases. Of these, the most
frequent suspected cause was wind blow (locally called nefas) (Figure 3). The
mechanism of causation was believed to be that the wind penetrates the chest as
one is exposed to it and then reaches the lungs causing TB.
60.7%
10.2%
1.0%
25.8%
0.5% 0.5% 1.0% 0.3%0%
10%20%30%40%50%60%70%
Cough
Haem
optysisFever
Chest pain
Loss of Appetite
weight loss
night sweating
Tirdness
,
Figure (2): Chief complaints of smear positive pulmonary TB patients in Amhara region.
5.2%
61.4%
6.3% 6.5%
17.5%
2.3%
0%10%20%30%40%
50%60%70%
Asthma
Nefas
Malaria
Bronchitis
TB No Dx
Figure 3 Perceived self-diagnosis among smear positive pulmonary TB patients in Amhara region
As to the first action taken during the onset of cough, 46.9% of the respondents
reported that they had tried self-treatment to cure their illness. They used
traditional homemade remedies to lessen their cough and enhance the smooth
expectoration of the sputum. The types of remedies used were, various plant and
animal products including steam inhalation (table 6). According to the
respondents, the duration of the treatment ranged from 5 to 8 days.
Table 6 Lists of local remedies used for treatment among smear positive pulmonary TB Patients in Amhara region________________________________________________________________________Types of treatment Number %________________________________________________________________________Hot fluids (atmit, suf, telba) 97 25.3 Honey, row egg yolk, Garlic 49 12.7Steam inhalation 28 7.2Goat meat (fat), ocholoni 14 3.6No treatment taken 204 53.1
3.3. Lengths and associated risk factors of the different delays
3.3.1. A. All health providers considered as a reference point
3.3.1. A.1 The choice of first health provider and the period of health seeking
After trial of home treatment and as the symptoms persisted, patients started
seeking health care from different health care providers. Of all the respondents,
61.7% initially visited non-formal health providers and 38.3% visited the formal
medical providers. The decision about where to go for help was influenced in the
majority (88%) of cases by close family members. Friends and health
professionals also took part in influencing 44 (11.3%) the patient. Almost all
patients had visited a health care provider in one-month time from the onset of
their symptoms. The type of the specific health provider visited during the initial
period of the illness is presented in figure 4.
Figure 4 Pie chart showing first health providers visited by smear positive pulmonary TB patients in Amhara region
27.10%
31%15.10%
3.90%
9.90%3.60%
9.40% Traditional providers
Drug retailoutlets
HC
Hospitals
PMP
Local injectors
Cli/HP
KeyPMP: private medical providers, CLi/HP: Clinic or Health Post, HC: Health Center
A general analysis was performed to assess the relationship of selected socio-
demographic variables with the choice of first health provider. For this purpose,
two groups were formed, Formal and non-formal health providers. Those who
first visited medical providers were categorized under formal and those who first
visited non-formal health providers were categorized under non-formal health
providers, and these groups were cross-tabulated against the socio-demographic
variables. The result showed that students (ORadj=0.32, 95%CI 0.11, 0.90) and
those with educational status of 9th grade and above (ORadj=0.42, 95% CI 0.23,
0.81) were less likely to visit the non-formal health providers. Otherwise, the
choice of health providers did not vary according to other variables (table 7)
Table 7 Relationship of choice of health providers with socio-demographic characteristic among smear positive pulmonary TB patients in Amhara region._____________________________________________________________________________
Visited formal visited non formal Crude AdjustedCharacteristics health provider health service OR (95%CI) OR (95%CI)_____________________________________________________________________________
Table 9 Choice of first health provider and its influence on the health providers’ delay among smear positive pulmonary TB patients in Amhara region._____________________________________________________________________
and those with income above 301 Birr (ORadj=0.39 95% CI 0.20, 0.76) seemed
to have lesser risk of increased health providers’ delay. However, when doing
multivariate (logistic regression analysis) these tendencies lost significance.
Table 10 The associations of socio- demographic and health service factors with health providers’ delay among smear positive pulmonary TB patients in Amhara region.
Delay No delay Crude AdjustedCharacteristics >62 days ≤61 days OR (95%CI) OR (95% CI)
Figure 7 Cumulative distribution of patient’s delay in relation to place of residence among smear-positive pulmonary TB patients in Amhara region.
P < .001
Logistic regression analysis was performed to look for the possible associations
of the different variables with significant patients’ delay (Table 12). Those who
lived beyond 10 Km radius of a medical facility (ORadj= 3.81, 95%CI 2.21-6.57),
age>45 years (ORadj=2.62, 95% CI 1.13-6.02) and self-treatment (ORadj=1.69,
95% CI 1.04-2.75) were significantly associated with increased patients’ delay.
We have also observed that the risk of patients’ delay among those who did not
visit non-formal health provider prior to visiting medical provider was smaller
(ORadj=0.34 95%CI 0.20, 0.57)
In univariat analysis those with education above 9 th grade (COR=0.30, 95%CI
0.17, 0.53), civil servants (COR=0.42 95%CI 0.22, 0.82), and students
(COR=0.35, 95%CI 0.16, 0.79) seem to have a lower risk of patients delay.
However, when we tried to find out the most influential factors of patients’ delay
using logistic regression analysis, these tendencies lost significance (Table 12).
Table 11 Sub-group analysis showing median patients’ delay among smear positive pulmonary TB patients in Amhara region______________________________________________________________________________
Median P-valueCharacteristics No patients’ delay (IQR) in days ________________________________________________________________________________Total n=384Sex
Male 202 30 (15-90) Female 182 35 (15-90) 0.27
Age 15-24 127 30 (15-60) 24-44 216 35 (15-90) 0.007 >45 41 60 (60-120) Marital status
Income No income 185 42 (15-90)Irregular income 114 30 (15-90)Regular income 36 21 (14-60) 0.16
1-300 Birr 30 (15-60)> 301Birr 49
HaemoptysisYes 96 67 (18-120) No 288 30 (15-71) <0.001
Table 12 The associations of socio-demographic and health service factors with delay to first visit to a medical provider among smear positive pulmonary TB patients in Amhara region.
The median delay from first visit to a medical provider to first reporting to a TBMU
was 15 days, mean 36 and IQR 0-53 days. The median medical providers’ delay
did not vary according to some selected socio-demographic factors. However,
those who first visited a health post or a private medical provider had longer
delay compared to those who visited a government health center (Kruskal-wallis
test; p< 0.001) (table 13).
Table 13 Sub-group analysis showing median medical providers’ delay among smear positive pulmonary TB patients in Amhara region________________________________________________________________________Variable N Median delay (IQR) P-Value________________________________________________________________________Sex
In this study, subjects were found to have visited a considerable number of
medical providers before ending up in the final diagnosis. The median number of
medical providers seen prior to starting TB treatment was 2 (IQR1-3) and the
highest was 8. This is not taking in to account the number of visits made to the
same medical providers. For 86 (22.5%) of the subjects, the total number of
medical provider seen exceeded 4 (figure 8). We compared the frequency of
medical providers visit prior to the diagnosis of TB with the socio-demographic
factors. There was no effect of sex, age, area of residence, education, income or
occupation on the number of medical providers seen (Table 14).
Figure 8 Graph showing the frequency of medical providers’ visit prior to the
diagnosis of TB among smear positive pulmonary TB patients in Amhara region,
September-December 2003.
Table 14 Comparisons of the socio-demographic factors with the frequency of medical providers visit prior to the diagnosis of TB among smear positive pulmonary TB patients in Amhara region._______________________________________________________________________Variable N Median number of (IQR) P-Value
of providers________________________________________________________________________Sex
Male 202 2 (1-3)Female 182 2 (1-3) 0.16*
Age15-24 127 2 (1-3)25-44 216 2 (1-3) 0.10*
>45 41 2 (2-4)Education
Illiterate 157 2 (1-3)1-8th 12 2 (1-3) 0.95*
>9th 82 2 (1-3)Distance
>10 km 168 2 (1-3)10Km 216 2 (1-4) 0.53*
Monthly income (Birr)No income 170 2 (1-4)Irregular income 104 2 (1-3) 0.25* *
Figure 10 Cumulative distribution of health system’ delay among smear-positive pulmonary TB patients in Amhara region.
Table 16 Selected socio-demographic factors associated with delay for more than three months before commencing anti-TB chemotherapy among smear positive pulmonary TB patients in Amhara region.________________________________________________________________________
Duration of illnessCharacteristics 120 days >121 days OR (95% CI)________________________________________________________________________Sex
Residence >10 Km 93 75 1:00 10 Km 170 46 0.33 (0.21, 0.52) *
IncomeNo income 120 65Irregular income 79 35 0.81 (0.49, 1.39)Regular income 25 11 0.81 (0.37, 1.76)
1-300Regular income 39 10 47 (0.22, 1.00)> 301
________________________________________________________________________* Significant at point < 0.05
Comparison among the various groups of respondents was made to look for
differences with regards to health systems’ delay (table 17). Significant
differences were not observed for most socio-demographic factors. But the
median health systems’ delay among the age group 15-24 was 14 days
compared to 27 days among the age group 25 and above years (Kruskal-Wallis
test; P=0.028) and those that visited private medical providers first had a longer
health systems’ delay compared to those that visited government medical
providers (P=0.022). The median health systems’ delay for those who went to a
health post was 39 days compared to 14 days in those that went to a health
center or hospital (Kruskal-Wallis test; P<0.001).
In logistics regression analysis, those who first visited a health post (ORadj=
3.50, 95% CI 1.86-6.57) or a private medical provider (ORadj=2.10 95% 1.18,
3.71) were significantly associated with increased health systems’ delay (Table
18).
Table 17 Sub-group analysis showing median health systems’ delay among smear positive pulmonary TB patients in Amhara region___________________________________________________________________________
Median health systems’ P-valueCharacteristics NO Delay (IQR), in days ___________________________________________________________________________Total n=384Sex
Male 202 20 (7-60)Female 182 22 (8-60) 0.59
Age 15-24 127 14 (6-43) 24-44 216 27 (8-43) 0.028 >45 41 27 (7-64) Marital status
Table 18 The associations of socio-demographic and health service factors with health systems’ delay among smear positive patients in Amhara region_________________________________________________________________________
Delay No delay Crude AdjustedCharacteristics >16 days ≤15 days OR (95%CI) OR (95% CI)
The subjects’ occupation appeared to influence the median total delay. Those
who reported shorter delay include self-employed 67 days (IQR, 38-112days),
civil servants 65 days (IQR, 35-98days) and students 51days (35-69days).
Where as longer delays were reported among housewives, 98 days (IQR, 65-172
days), farmers 97 days (IQR, 61, 188 days) and by the unemployed 78 days
(IQR, 45-155 days). It also varied with the patients’ educational level. Illiterates
had a median total delay of 98 days (IQR 65-186 days) and those with education
above the level of 9th grade, the median total delay was shorter, 48 days (IQR 34-
83 days) (Kruskal-Wallis test; P=0.005).
The subjects’ marital status was found to affect the median total delay. Married
subjects were found to have the longest delay, 97days (IQR 62-172days)
compared to those who were never married, 67 days (IQR 36-126 days)
(Kruskal-Wallis test; P=0.019). Age also had an effect on the median total delay.
Those who were between 25 and 44 had longer delay compared to those below
24 years of age (P=0.016).
Patients who treated themselves at the onset of the initial symptoms at their
home had a longer total delay compared to those that did not try self-treatment.
The median total delay was 95 days (IQR 65-185 days) vs 65 days (IQR 36-
123days) (Mann-Whitney test; P=0.005). Also, patients with haemoptysis had
longer duration of total delay than those with out haemoptysis (median 125 vs
67days Mann-Whitney test; P=0.005)
Table 20 shows the association of socio-demographic and health services factors
with the median total delay. Accordingly the age group 25-44 (ORadj=1.85,
95%CI 1.06, 3.20) was significantly related with increased total delay compared
to the age group 15-24years. Marital and educational statuses of the study
groups were associated with increased total delay. The married group had
increased risk of delay compared to the singles (ORadj= 2.26 95% CI 1.14, 4.10)
and those with education above the level of 9th grade had a smaller risk of total
delay compared to the illiterates (ORadj= 0.42 95%CI 0.19, 0.89). Besides these,
those who first treated themselves (ORadj=1.75, 95%CI 1.05, 2.93) and those
who went to the non-formal health providers (ORadj =2.52 95%CI 1.49, 4.23)
were characterized as having a higher risk of longer median total delay.
Regarding the knowledge of patients about TB, it was found that those with low
level of knowledge (ORadj=3.49 95% CI 2.01, 5.80) had a 3-fold risk of total
delay than those having lower level of knowledge about TB.
In univariat analysis age >45 years (ORadj=1.67 95%CI, 0.72, 3.80), civil
servants (OR=0.39, 95%CI 0.20, 0.76), students (OR=0.21, 95%CI 0.87, 0.50)
and self-employed (OR=047, 95%CI 0.23, 0.96) seem to have a smaller risk of
delay compared to the farmers. But when we analyzed the possible effects of the
interaction of the entire variables using multi-variat (logistic regression) analysis,
these tendencies lost significance.
Table 19 Sub-group analysis showing median total delay among smear positive pulmonary TB patients in Amhara region___________________________________________________________________________
Median totalCharacters tics No Delay (IQR) P-value ___________________________________________________________________________Total n=384Sex
Male 202 69 (38-125)Female 182 93 (61-150) 0.095
Age15-24 49 64 (35-100)
25-44 278 94 (62-147) 0.016 >45 57 95 (97-194)Marital status
_____________________________________________________________________________Table 20 The relationships of socio-demographic and health service factors with late initiation of anti TB chemotherapy among smear positive pulmonary TB patients in Amhara region.______________________________________________________________________
Delay No delay Crude AdjustedVariable > 81 days ≤80 days OR (95% CI) OR (95% CI)_______________________________________________________________________Sex Male 91 111 1:00 1:00 Female 83 99 1.33 (0.89, 1.99) 0.81 (0.48, 1.36)Age 15-24 49 78 1:00 1:00 25-44 118 98 1.92 (1.23, 2.99) * 1.85 (1.06, 3.20) *
In general, taking all health providers and only medical providers as a
reference point separately, we can see variations in the length of the different
delay periods. In the first scenario (figure12) patients who first visited a
medical provider had a median delay of 30 days. In the second scenario
(figure 14) considering all health providers as potential venues of health
seeking, we can see that patients visit a health provider relatively early
compared to the first scenario. The health-seeking period is shorter. It took
them only 15 days to first visit a health provider. Overall the contribution of the
health-seeking period by patients to the total delay is smaller. As shown in
figure 14, the greater portion (81%) of the delay was due to the health
provider.
30
21
80
0
10
20
30
40
50
60
70
80
90
Patients’ delay Healthsystems’delay
Total delay
med
ian
in d
ays
Figure 12 Graph showing the median delay periods taking only medical providers as a reference point among smear positive pulmonary TB patients in Amhara region.
15
62
5
80
0
10
20
30
40
50
60
70
80
90
Health seekingperiod
Healthproviders’ delay
Diagnosingfacility delay
Total delay
med
ian
in d
ays
Figure 13 Graph showing the median delay periods taking all health providers as a reference point among smear positive pulmonary TB patients in Amhara region.
Finally, as a summary we have presented a diagram showing the different delay
periods starting from onset of cough until the patient is put on anti-TB
chemotherapy. It also shows the patients’ possible choices of health providers.
(See figure 15)
3.4 TB diagnosis at the private medical providers
In this study, it was found that 92 (24%) of the total study subjects were
diagnosed and referred to the TBMU for treatment by private medical providers.
According to the national TB control program guideline of Ethiopia, all smear
positive pulmonary cases diagnosed at the private medical providers must be
referred to the government TBMU for initiation of chemotherapy. But in this study,
it was found that 14 (15.2%) of the subjects after being diagnosed at the private
doctors had prescriptions of anti-TB drugs. These patients directly started anti-TB
chemotherapy by purchasing drugs at the private pharmacies. During the
interview the subjects reported that, they took anti-TB drugs for 1-2 month. The
18.70%
81.30%
Health seekingperiodHealth providers’delay
Figure 14 Pie chart showing the contribution of health seeking period and health providers’ delay to the total delay among smear positive pulmonary TB cases in Amhara region
reason that they came to the TBMU was that the pharmacies had run out of
drugs.
3.5 Stigma
TB was found to be stigmatized among the respondents. Two hundred eighty
five (75%) believed that TB is a social stigma and 270 (70%) responded that they
would not enter others social circle for fear of not being accepted by others.
Subjects also closely linked TB and HIV. Two hundred and five (53.4%) of the
respondents said that TB is associated with HIV and of these, 179 (47%)
believed that going to a medical provider for TB test can make other people think
that the person has AIDS. Forty-eight (12.5%) said that they had a fear of being
tested for HIV when they initially reported to the medical provider. However, we
did not find associations between stigmatizing attitudes to wards TB/HIV and
patients’ delay.
Chapter four: Discussion
4.1 The distribution of the sample population
In our study, the distribution of the study population by sex showed that the
proportion of males exceeds that of the females. This is quite similar with the
notification trend at national and regional levels (11). It is also similar with the
global trend. According to Rieder (1999) in virtually all countries, notification rates
among males are higher than among females (34). The distribution of the
population by age also showed that the majority was in the productive age group.
The mean age in the study population was 29 years. This is also similar with the
trend in general in developing countries. In developing countries, TB peaks in
young adults and it is estimated that 75% of the TB cases notified are in their
productive age group (71). The higher proportion of the Amhara ethnic group
and Christians in the sample population follows the general pattern of population
distribution for the region (12).
In this study, 40.9% of our sample population were illiterate (especially females
being more illiterate), had no defined income and were sharing a single room
with an average family size of five. This clearly indicates a very poor socio-
economic condition among the study population. As Rieder (1999) described it,
TB and poverty are strongly associated and low socio-economic indicators tend
to result in conditions that are conducive to increased transmission of tubercle
bacilli, resulting thus in a generally higher prevalence of TB infection with
subsequent increased incidence of the disease. According to the World Bank
classification, 78% of the 22 countries with the highest TB burden in the world
are low-income countries and it is known that Ethiopia is in this group (13). In
general, the current TB situation in this sample population clearly confirms the
fact that TB is a disease of lower income and lower resource countries.
4.2.1 The health seeking period and health providers’ delay
In this study, we were very much surprised to observe that many of our subjects
visited the health providers to seek for health care quite early. The delay period
from onset of major symptoms to first visit to a health provider was only 15 days.
This is in fact the same as the result documented in the Gambian study (19),
which clearly indicates that, patients seek health care early but the type of health
providers they visit varies considerably. On the other hand, 81% of the delay was
attributed to the health providers. The median health providers’ delay was 61
days. This is higher than the result reported in the Gambia (19). The long health
providers’ delay in our study might be related to the fact that the majority, 61.7%,
of the subjects initially went to the non-formal health providers to seek for health
care, in which case the likelihood of being referred to the formal health providers
for diagnosis might be less. As a result, patients might spend longer time before
they get the correct diagnosis.
The logistic regression analysis performed to analyze the possible association of
the factors between the socio-demographic factors and the median health
providers’ delay showed that the literates in general, those who lived within 10
Km radius of a medical facility and those who had visited the formal health
providers initially had lower risk of increased health providers’ delay. This might
be related to the fact that being literate, subjects might have better knowledge of
TB so that they might frequently visit health providers till they get diagnosed.
Residing within 10 Km radius of a medical facility is also an advantage to seek
for health care and of course visiting a medical provider initially might help to get
the diagnosis in a relatively shorter period of time than visiting the non-formal
health providers. On the other hand, those who were between 25-44 years of age
had experienced longer health providers’ delay. The reasons might be related to
the fact that these are within the productive age group. They are the working
force in the community. Much of their time is usually allocated for work. As a
result they might fail to give priority for health care during their illness. They might
also choose the shortest option i.e. buying drugs from the drug retail outlets to
save time. This is also evidenced in this study by 31% of the respondents
primarily visiting the drug retail outlets at the onset of the present illness.
4.2.2 Patients’ delay
Taking only medical providers as a reference point, we found a median patients’
delay of 30 days. This finding is similar with other studies that were conducted in
Botswana (18), Ghana (16), Philippines (73) and Penang (74) that showed a
median patients’ delay of 3 - 4 weeks. On the other hand, our finding is much
lower compared to the previous two Ethiopian studies that showed median
patients’ delay of 179 and 60 days, respectively (22, 23).
The reasons for our relatively shorter patients’ delay compared to the previous
Ethiopian studies might be related to the fact that we considered all levels of
health care including the lowest health care facilities (clinics & health posts) as
relevant sites for first health seeking contact for our study subjects. This may
have shortened the patients’ delay, as health posts and clinics are relatively
closer to the community compared to the health centers and hospital, which are
concentrated in major towns. Health centers and hospitals are diagnosing
facilities in the region. The previous two Ethiopian studies considered subjects
coming to these facilities only.
This relatively short patients’ delay may also further be explained by the fact that
currently there is a rapid change of health care system in the country, where
there is an increase in the participation of the private sector and expansion of the
health service to the population (11). As a result of these facts, patients might
have a better access to health care than before. Hence this may have resulted in
a short duration of patients’ delay. Therefore, we should not be surprised at
observing longer patients’ delay in the two previous Ethiopian studies that
considered health centers and hospitals as the lowest level of health care contact
for patients, in which case, the likelihood of presenting late may be very high.
Because, these facilities were inaccessible for the majority of the population in
the previous times when the studies were conducted
On the other hand, even though our median patients’ delay seemed to be
shorter, 48% of the subjects were delayed for more than 30 days prior to
presenting to a medical provider. This delay was significantly associated with
older age and distance from a health care facility. These have proved to be also
important factors in other studies that were conducted in Zambia (48) and south
Ethiopia (22). In our case, the reasons might be related to the fact that old people
are usually dependent on other persons which makes it difficult for them to visit
health facilities early. There is also generally poor access to health care for the
regions’ population.
Low level of knowledge about the symptom and treatment of TB was associated
with patients’ delay. This is in accordance with the previous study done in
Ethiopia (23), Vietnam (58) and Tanzania (50). Patients who presented with
haemoptysis had also a longer patients’ delay. In this regard, similar observation
was made in south Ethiopia (22), which suggests that patients stay at home until
they observe an alarming symptom like haemoptysis.
We observed a significant association between self-treatment at home and
longer patients’ delay. Studies conducted in other African countries such as
Ghana (16), Botswana (18) and Kenya (53) also showed the same result. Self-
treatment was also found to be a common practice in the previous unpublished
Ethiopian study (57). Patients usually start treating themselves with homemade
remedies during their early symptoms. It is when the symptoms get worse that
they start seeking help from medical providers. This may be related to poor
knowledge of TB symptoms and its treatment among the population. It may also
be related to poor access of medical care to the general population.
In the present study, we did not find an association between educational status
and patients’ delay and between occupational statuses and patients’ delay.
However, the study from south Ethiopia has documented that being illiterate, a
house wife and a farmer were associated with longer patients’ delay (22). In our
study, even though these factors seemed to be associated in the univariat
analysis, their effect disappeared in the multivariate (logistic regression analysis).
This indicates that education and occupational statuses did not turn out to be
potential predictors of patient’ delay.
Other studies have documented that females took longer time to seek for health
care compared to males (16, 58). This finding could not be confirmed in our
study. In our case, the presence of integrated health service at all levels of health
care that gives more emphasis to women and children, the continuous
campaigns on polio, measles and tetanus immunizations might have helped the
women to appreciate the benefit of medical care. Mothers usually visit health
care facilities for ante-natal care, family panning and for immunization service for
their children. During these times they attend health education sessions in health
institutions. This might have an effect on their health seeking behavior. It seems
also that women in our case do not have decision making problem for seeking
health care. As indicated in the results section, 85% of the married subjects
claimed that they could decide on their own about where to go for help during
their illness. Therefore, all these reasons in our case might have helped the
women to seek for health to the same extent as men.
TB and HIV seem to be closely linked in the peoples’ mind in Ethiopia. The
current study showed that 47% of the subjects believed that TB and HIV are
associated and said that, coming to a health care facility for TB symptoms can
make other people think that the person has AIDS. Other Studies in this regard
have shown that, there is discrimination that surrounds HIV and TB patients that
may prevent them from seeking health care by going to public health care
facilities (64). However, the current study was not able to demonstrate a
significant relationship between those who expressed their feeling about the
stigmatizing attitudes towards TB and HIV and delay to coming to medical
providers. This might be related to lack of openness among the subjects in
expressing their genuine feeling about stigmatizing attitudes during the interviews
or it might also be a true finding.
4.2.3 Health systems’ delay
With regard to the health systems’ delay, our result showed a median delay of 21
days. This is more or less similar with other studies conducted in Tanzania (46),
Penang (74), New York (47) and Japan (75) that showed a median health
services delay of 3 weeks to 1 month. On the other hand, a relatively shorter
health service delay of 6 days was observed in the previous Ethiopian study (23).
Our result shows a longer delay compared to this study. This may be related to
the study setting in which case the previous Ethiopian study (23) included
subjects who presented initially to a diagnosing facility, which makes it of course
shorter as these patient can be evaluated and diagnosed on the spot. Whereas
our study population included patients that were referred from clinics, health
posts and private medical providers without diagnostic facilities. This might make
the period of diagnosis and commencement of treatment longer as it takes
considerable period of time to reach the TBMU from the time of referral.
In this study, prior attendance to a health post/clinic was a risk factor for longer
health systems’ delay. Similar finding was observed in Botswana (18). Our
finding may be explained by the fact that, health posts and clinics are run by
health assistants and junior nurses whose primary training is not to diagnose
serious diseases but to concentrate on patient care and preventive activities.
Besides this, these facilities are not equipped with the necessary diagnostic
equipments like microscopes and others. As a result, subjects might be
misdiagnosed and mismanaged. This might cause prolonged delay before
diagnosis. Prior attendance to non-formal health providers was also significantly
associated with longer health systems’ delay. This finding is a little higher than
the results documented in the study done in Dabat district in Northwest Ethiopia
(76). The common reasons for not visiting formal health providers in our study
were that, illness was considered harmless followed by health institution being
very far from home and the feeling that self-treatment was sufficient. These
responses could be related to several issues including lack of knowledge of
symptoms of TB, the perception of the relatives on how long after the onset of an
illness it is proper for a patient to still go to health institutions, lack of confidence
in modern health care and poor access to medical providers.
In the current study, those who visited private medical providers had longer
health systems’ delay. Similar finding was observed in Penang (74), where
patients who first consulted a private practitioner were the least likely to be
diagnosed appropriately. The reason for long health systems’ delay in our case
might be related to the guideline of the TB control program of the Regional
Health Bureau that strictly forbids the private medical providers to treat TB cases.
According to the guideline (12), patients diagnosed as having smear positive
pulmonary TB must be referred to the government TBMU for the initiation of anti-
TB chemotherapy. Patients will be re-examined and sputum for AFB will be
requested for the second time at the TBMU. Therefore, there is no doubt that all
these procedure might take additional time for patients who visited the private
medical providers prior to diagnosis and starting treatment compared to other
patients that directly went to the TBMU.
Apart from the above significant risk factors, we did not find associations
between socio-demographic factors and longer health systems’ delays. This
finding is similar with other studies conducted in countries like Zambia (64),
Botswana (18) and Penang (74) but different from Ghana (16) where women and
rural residents had longer health systems’ delay and a previous Ethiopian study
(23) in which distance to health care facility was found to be a risk factor.
With regard to medical providers’ delay, long distance to the TBMU, being
illiterate and prior attendance to a health post were risk factors for being delayed
for more than 3 months in some of our respondents. This may be related to the
fact that diagnosing facilities are found in major towns. Therefore, for those living
in rural areas these facilities might not be easily accessible, as a result, these
patients might delay longer before reaching diagnosing facilities. Patients who
are illiterate might not also go to the higher level even though they are referred.
They might not take the referral seriously. Because the general knowledge about
the consequences of late presentation might be low. Therefore, they might not be
motivated to go to the next level of health care earlier compared to the literates.
With regard to the relation between prior visits to health posts and medical
providers’ delay, it might be explained by the fact that, health posts are not well
equipped with well-qualified health professionals and equipment to diagnose TB
compared to health centers and hospitals. Therefore, patients might be
misdiagnosed or mismanaged resulting in longer delays.
4.2.4 Diagnosing facilities’ delay
As for the diagnosing facility’s delay, the present study showed a median
diagnosing facilities’ delay of 5 days. This is a little bit higher than the Gambian
study that showed a median diagnosing facilities’ delay of 0.2 weeks (22). This
may be because of the fact that the Gambia is a small densely populated country
with 87% of the population having a good access to health care within a 3 Km
radius (19), unlike Ethiopia where the health service coverage is not beyond 50%
(3). But still, we have a feeling that, the diagnosing facilities’ delay observed in
this study is not that wide. We would say that patients get their diagnosis and
commence their treatment within a reasonable period of time as long as they
manage to reach the TBMU.
Patients in this study were shopping for treatment for a considerable period of
time prior to diagnosis. The average (mean) number of medical providers visited
was two. This is excluding repeated visits made to the same providers. Our
finding is different from the Gambia where patients visited 4 medical providers
prior to diagnosis. This might show that our patients are referred earlier to the
TBMU or patients might not frequently shop for treatment from different medical
providers, as the numbers of private providers are relatively small compared to
the Gambian study, or patients might stick all the time to the first private medical
provider till diagnosis or referral.
4.2.5 Total delay
The median total delay observed in this study was 80 days. This is different from
the previous south Ethiopian study (22) that showed a median total delay of four
months. However, a more or less similar finding was observed in other countries
like Gambia (19) Botswana (18), and Penang (74) that showed a median total
delay of 8.6 and 12 weeks, respectively. The reason for our relatively shorter
total delay compared to the previous south Ethiopian study may be related to the
increase in DOTS coverage in the region. Currently, 51% of the region has been
covered by DOTS unlike the previous years where DOTS coverage was very low
(12).
The risk of increased total delay was also higher in those with education less
than 9th grade and in married couples compared to those above 9th grade and the
singles, respectively. In this regard, similar finding was observed in the previous
south Ethiopian study (19) and may be explained by the fact that subjects with
lower level of education might have poor awareness regarding the symptoms and
treatment of TB and married couples might have shortage of time to care or give
attention for themselves, as they are usually responsible for the entire family.
In this study, far distance to patients’ home and low-level knowledge about the
symptoms and treatment of TB were also associated with increased median total
delay. This result is consistent with the findings in Botswana (18), southern
Ethiopia (22) and Vietnam (58). As has been described in the results section, the
majority (57 %), of the patients at the onset of their illness related their symptoms
to other diseases. Moreover, most of the study subjects believed that, the
symptoms would disappear by themselves. This clearly shows a lack of
knowledge among our study subjects with perceptions and practices that might
delay the patient. Therefore, we were not surprised to observe this significant
association in our study. The same might be true for the association observed
between self-treatment and increased total delay.
In general, we have seen from the above discussion that delay in diagnosis has
been studied in many countries. In all of these studies, a range of conflicting
differences on the lengths and risk factors of diagnostic delays have been
documented. Likewise, the present study showed similarities and differences in
the choice of first health providers, and in the lengths and risk factors of the
different delays when compared to other studies. This may due to various factors
related to the study setting, the method of estimating the time from onset of
symptoms to seeking medical care and the characteristics of the study population
under which our study was conducted.
With regard to the choice of first health provider, in almost all the previous
studies reviewed, only first contact with a medical provider was taken as a
reference point when analyzing delays. However, one study from the Gambia
(19) incorporated all health providers in this regard. In the current study, all
health providers were considered as a reference point for analysis. This is
because, in Ethiopia, all health providers are assumed to be potential sources of
health care for TB patients.
4.3 The role of knowledge, perception and behavior in diagnostic delay
In the health seeking behavior model, the process of care seeking begins with
“symptoms recognition” (77). Likewise, in the present study, we observed an
alternative explanation for the symptoms of TB among the respondents. In the
initial period of the onset of cough, 61.4% of the respondents attributed their
cough to other conditions like nefas (bird). Nefas/bird in Amharic simply means
wind blown to the chest. It is also seen in relation to exposure to the cold (any
cold element, like the cold water). It is a traditional phenomenon but not a
scientifically explained disease. Similar finding was observed in a recent
unpublished qualitative study that was conducted in Addis Ababa the capital of
Ethiopia, which showed that the first symptom of TB (cough) was often not
directly related to TB. The patients’ belief was that they had bird (57). Other
studies conducted in Colombia, attributed TB symptoms to flue (52). In Malawi
TB symptoms were attributed to tsempo, mphumu or modulo which is a
traditional folk condition believed to be caused by careless sexual behavior or
adultery (54).
In Ethiopia, there is a strong belief that nefas /bird causes samba nekersa
(pulmonary TB). Especially a narrowly opened window (while traveling in a car or
sitting in a room) is believed to be dangerous as a sharp wind could come
through it with a power enough to penetrate the chest and then directly to the
lung causing TB. It is a common phenomenon to observe many people taking
preventive action against nefas/bird. For example, when people are traveling by
a taxi or a bus, windows are usually closed. My own experience confirms this. As
I was frequently traveling for supervision purpose during the data collection
period of this study, I some times used public transport like bus. In the bus, we
were sixty traveling 470 Km together. From the time we departed till we reached
at our destination, all the windows were closed. As I suffered from sever
headache and suffocation, I tried to open the window once, every body shouted
at me saying that, "You guy! nefas is coming! Shut the window! Please! Please!"
then I immediately closed. The only time we got fresh air was when the bus
stopped for breakfast and lunchtime. This clearly shows that, the wrong
perception among the people is causing them to behave in the wrong way. All
this may be related to lack of awareness regarding the causes of TB among the
population.
According to the health seeking behavior model, the first step in symptoms
recognition includes also identifying the causes and the severity of symptoms
(77), in the present study, we found that patients with haemoptysis had longer
duration of illness compared to those with out haemoptysis. It seems that
patients wait for long time until they recognize sever symptoms like
(haemoptysis). The reason for this may be related to the subject’s perception of
severity of diseases. Patients may assume that, when the cough gets severe, like
accompanied with haemoptysis or dyspnea, it is an indicator that the nefas/bird
has penetrated the lung, at this time they may be motivated to take one step
forward in the health seeking behavior, i.e. they might immediately go to a
medical facility. As severe weight loss and severe cough were strongly related
with TB in other countries (52), haemoptysis might have been considered as the
major sign of TB not cough alone by some of our respondents.
Following the health seeking behavior model (77), the next step was consulting
the symptoms with laypersons and making a decision about treatment. Likewise,
our result showed that the majority (88%) of the subjects consulted family
members regarding what to do in the next step. They also decided to use self-
treatment (home made) like variety of hot fluids and steam inhalation. The
practice of self-treatment in response to the initial symptoms is also evidenced in
the recent un published Ethiopian study where patients were taking different
homemade remedies to relieve their cough (57). Studies conducted in other
African countries such as Ghana (16), Botswana (18) and Kenya (49) showed
also the same result. Other studies from Pakistan (74) and Malawi (62) reported
that beliefs about the etiology of TB were associated with the health seeking
behavior. For example, people who believe that TB is caused by supernatural
forces would seek care from folk and traditional healer (62). According to
Kleinman, this seems a universal phenomenon in which case the response to
early symptoms of disease as well as their action upon it is within “the popular
sector” of the health care system (77).
Kleinmans (1980) described health care as a local cultural system consisting of
three overlapping parts; the popular, the professional and the folk sector. The
popular sector is the largest part of the system, consisting of a matrix containing
several levels; the individual level, the family level the social network and the
communities’ beliefs and activities. It is within this sector that, illness is first
defined and different health care activities initiated. It is also within this sector
that a large percent of illness episodes are managed. Self treatment by the
individual and his or her family is the first therapeutic intervention people make
use of in a wide range of cultures, and when people turn to folk medicine, and/or
modern western medicine, their choices are often based on the belief and the
value orientation of the popular sector (77).
Generally, from the above explanation, we can see that, initially patients had
wrong perceptions regarding the causes of TB. As a result, they attributed their
illness to other diseases and stayed at home taking self-treatment. Some of them
even stayed till they noticed an alarming symptom like haemoptysis. This has
resulted in a considerable delay before diagnosis and was clearly shown in
findings section by the association between self-treatment and patients’ delay
being significant.
4.3. The contribution of the different health providers in diagnostic
delay
4.3.1 Drug retail outlets
In this study, drug retail outlets were very much utilized among TB patients in the
early period of their illness. The result showed that 31% of the study subjects
used self-medication at the onset of their symptoms. Buying drugs from drug
retail outlets might indicate a preference for the more convenient way of getting
medicine. Previous studies from Northwest and southern Ethiopia reported a
higher prevalence of self-medication with a proportion of 24 and 27.6%,
respectively (79). Other studies in Mexico (75), India (80) and China (81) showed
30%, 34.5% and 32.5% prevalence of self-medication, respectively. However,
the current study might not be directly comparable with these studies as one
disease entity is investigated while the other studies considered all diseases
symptoms. On the other hand, the major symptoms that led patients to self-
medications in those studies were headache, fever and cough which are also
often symptoms experienced by TB patients
For self-medication, the study subjects reported that they used antibiotics like
ampicillin capsules and penicillin injections and said that they got improvement
with their cough after taking these drugs. Other studies (62) have noted that
respiratory symptoms in patients with microbiological confirmed TB can
temporarily subside after a course of antibiotics. This may be due to some
antibiotics having a short mycobacteriostatic action or because of bacterial super
infection. Whatever the reason may be, symptomatic improvement after a course
of antibiotics may contribute to a delay in diagnosis as has been found in other
parts of sub Saharan African countries (62).
Currently in Ethiopia, following the new health sector reform, quite a number of
drug retail outlets have been opened (3). Therefore, it is relatively easier to buy
drugs from drug retail outlets rather than going to a medical facility, which in fact
may cost additional money including transportation cost. One can find different
kinds of antibiotics in the drug retail outlets. Sometimes, even anti-TB drugs
might also be found (12) indicating the common practice of selling drugs without
prescription. Even though there is a law that forbids the selling of such drugs
without prescription. It seems that the guideline of the regional health bureau is
not being respected. As WHO noted, self-medication provides a cheep
alternative for people who cannot afford to pay a medical practitioner. Thus self-
medication is often the first response to illness among people with low income
(83) and all these might contribute to patients’ delay.
4.3.2 Traditional health care providers
Traditional health care providers were among the sources of health care visited
by the study subjects. The findings showed that 27% of the subjects initially
visited traditional health care providers. This is nearly similar to the finding of the
study done in Malawi (62) where 30 % of the study subjects initially visited
traditional health care providers, but lower from Tanzania where the proportion of
first visit to traditional health providers was 39% (50).
In Ethiopia, modern health service utilization appears to be generally low. An
earlier study which summarized the health profile of 52 districts reported that the
per capita annual number of visits was 0.23 visits over all, with the mean for
urban double that of the rural district (84). The vast majority of the Ethiopian
population, therefore, still depends on traditional medicine and its practitioners.
The 1982- 1983 rural health survey revealed that more than half of the health
service seekers relied on traditional healers. In Addis Ababa, which is one of the
highly urbanized centers of the country and where modern health services are
relatively accessible, 26% of the representative populations were shown to have
used traditional medicine (9). According to the 1999 World Bank report, more
than 80 % of the people use herbal remedies as their first choice for the day
today health care needs (85).
Traditional medicine remains of paramount importance to the Ethiopian people
and
Commands a great deal of acceptance among the majority of the population. The
widespread use of traditional medicine in the country among both rural and urban
population could be attributed to the following major factors (9).
1. Acceptability: Use of medicinal plants constitutes part of the cultural
heritage representing the identity and the uniqueness of our society.
Traditional medicine is a component of the local culture. People resort to it
even when there is demonstrably better alternative care.
2 Accessibility and affordability, as compared to modern drug. Medicinal
plants are often within easy reach and affordable both in terms of financial
resources and time when compared to modern drugs dispensed in
remotely located health institutions
Moreover, in Ethiopia, it is a common phenomenon that patients who have visited
traditional health providers normally do not seek formal health care from medical
providers till they finish the ordered herb/other remedy by the traditional health
provider for a given period of time. There is also a cultural belief that, the
herb/other remedy given by the traditional health provider would not work if one
took injections or tablets from the modern medicine at the same time. This and
the above mentioned reasons might delay patients for a considerable period of
time before seeking health care from the formal health providers for diagnosis.
4.3.3 Private medical providers
In the present study, 24% of the subjects had their diagnosis from the private
medical providers and were significantly delayed compared to those directly
diagnosed in the TBMU. Even though TB treatment is offered exclusively in the
public sector, many studies have revealed that patients tend to be under private
care for a considerable length of time before TB is diagnosed and patients are
referred to the TB service. For example, in Sao Paulo city (86) where TB care
largely takes place in the public sector, an analysis of the place of first diagnosis
and the extent of delay in diagnosis showed that in about 20% cases the
diagnosis was first made in the private sector. The mean delay in diagnosis was
12.5 weeks. In the Kenyan study (88), 90 % of TB suspects claimed that they
had attended private health care facility yet 65% had neither a chest radiograph
taken nor their sputum examined. A study of TB patients and practitioners in
private clinics in India (89) showed median delay in diagnosis of about 2 to 3
weeks among urban and rural patients after they sought help at private clinics.
About 33% of the urban patients and 36% of the rural patients had not been
diagnosed even after 4 weeks of seeking help. Another study in Vietnam (89)
showed that patients who had first turned up to a private physician were more
likely to have a long provider delay compared to people who had first turned to
the national TB control program. This might show that private providers follow
poor diagnostic practices leading to long delays in diagnosis.
In the present study, it was also revealed among some patients that, private
medical providers instead of referring suspected/confirmed cases of pulmonary
TB to the TBMU, they prescribed anti-TB drugs and informed the patient to
directly purchase the drugs from private pharmacies. This practice is clearly
against the guideline of the national TB control programs, which indicates that
there might be poor control of the private medical providers by the district, zonal
or regional health bureau. The availability of anti-TB drugs in the private
pharmacies might lead to irrational use of drugs, which ultimately can result in
the emergence of drug resistant TB in the region that will make the TB control
program unsuccessful (12).
4.3.4 Local injectors The study showed that considerable number (3.6%) of the subjects visited local
injectors during the onset of their illness. According to the ANRSHB (2002) there
exists a practice both among the private practitioners as well as the local
injectors to give injections like penicillin, vitamins and saline water for patients
with nefas/bird. Similar observation was made in the previous unpublished
Ethiopian study where saline water was extensively given as injections (57). The
use of injection is also high in other sub-Saharan African, Middle East and South
East Asian countries by the informal private providers (90).
Local injectors in Ethiopia are part of the health care system both in the urban
and rural areas. They may or may not have medical education but they
administer injections to patients presenting to them. They are not legally allowed
to practice as a local injector in any way. The procedure is usually performed
behind closed doors.
Ethiopian patients especially in rural areas strongly prefer injections to tablets.
This may be related to the belief prevailing among the population that, “injections
radically remove the disease compared to oral medications.” Patients might
understand the pain that they feel during the injection procedure as a sign that
the disease is gone forever. Many patients are dissatisfied if medications are not
given while diagnostic tests are pending or the illness does not necessarily call
for medication (91). Therefore, due to the strong belief in injection among the
population, considerable numbers of people might go to local injectors for their
illness. As a result a placebo effect may cause delay in diagnosis among TB
patients.
Strengths of the study
1. Relatively large sample size was taken
2. We were able to meet the intended sample size within the study period
3. Very close supervision was conducted and the data collection activity was
successfully accomplished
4. We did additional analysis on other variables besides our original
objectives
Weaknesses of the study
1. We were not able to collect data on sputum grading. Because, most of the
TBMUs were not doing it during the study period. Therefore, we were not be able
to analyze the infectivity versus the duration of illness
Limitations of the study
There are limitations to our study. One of the limitations is related to the
interviewers. We used health professionals to conduct the interviews. In this
regard, we have a suspicion that subjects might underestimate the duration of
stay at TBMUs for fear of being mismanaged by health professionals during
subsequent visits.
The second one is related to selection bias. We included patients presenting to
government health care facilities. Other pulmonary TB patients who might
probably go to private medical providers and who stayed at home during the
study period were not included in our study. Here it might be difficult to consider
our study result as representative of all smear positive TB patients at national or
regional level. Because, the nature and behavior of patients that were not
included in our study might vary. However, we have a strong belief that, our
sample is representative of the smear positive pulmonary TB patients presenting
to TBMUs in Amhara region as we managed to select the representative health
facilities. We were also generally able to interview 35% of the total diagnosed
cases in the region during the study period.
The other problem is related with the recall of the duration of illness. We
interviewed our subjects retrospectively to tell us what happened during the initial
period of their illness. Therefore, when we see it from the angle of validating the
duration of illness, it may suffer from recall bias in some of the respondents as
they might not be able to tell us the exact date of onset of their illness. However,
we have put in our maximum efforts to minimize this problem. We specifically
asked the onset of the major symptoms and how long after these symptoms they
consulted a health provider. As described earlier 96.4% had cough that is likely
to be remembered by the subjects. Moreover to estimate the date of onset of
symptoms, we have used local calendar listing the main religious and national
days. As Christianity is the predominant religion in the study area, it is expected
that the people might remember these days. As shown earlier in our results
section, the length of recall period for the majority (69%) was below 2 months.
Therefore, we might say that, the possibility of recall bias is very much reduced
though it is admittedly difficult to eliminate it altogether.
As a data collection tool, we used structured questionnaires. We had also some
semi-structured questionnaires. We believe that the tools were appropriate in
gathering the information for our research questions. These methods are the
most commonly employed methods used to gather information for cross-
sectional studies where the numbers of respondents are usually large (92).
However, there are some disadvantages of using these tools. Like for example,
the interviewer may inadvertently influence the respondents, important
information may be missed, because spontaneous remark by respondents is
usually not recorded or explored, and open-ended questionnaires are difficult to
analyze. To control these problems, all the interviewers were instructed to take
all the necessary care during interviewing to make the respondents as free as
possible with regard to responding as fully as they felt, the interviewers
comprised also both males and females (30% of the interviewers were females).
All were from the same ethnic group who were well aware of the socio-cultural
issues in the study area, they were also local residents who spoke the native
language, and in analyzing the open ended questions, after the data were
collected, we went through all the possible responses and carefully categorized
them for analysis.
Chapter 5: Conclusions & Recommendations
Generally, our study showed that there was a significant delay in the initiation of
anti-TB chemotherapy among smear positive pulmonary TB cases in Amhara
region. This delay was to a large extent attributed to the long health providers’
delay. Eighty one percent of the delay was due to the health providers, and the
overwhelming majority (61.7 %) of pulmonary TB patients in the region attended
non-formal health providers as their first preference when symptoms initially
started.
To our surprise we found that patients seek health care relatively early and the
type of health provider they visit vary greatly. By saying this, we are not totally
denying the contribution of the patients’ delay to the total delay, as 48% of the
subjects took more than 30 days prior to reporting to the medical providers. The
major factors associated with the patients’ delay were related to the wrong
perception regarding the causes and symptoms of TB and related behaviors
(self-treatment), lower access to medical providers and prior attendance to the
non-formal health providers. On the other hand, the major factors associated with
the health systems’ delay were prior attendance to the health posts and private
medical providers. These were the potential risk factors affecting patients’ and
health systems’ delay in the diagnosis of pulmonary TB in Amhara region. The
median diagnosing facilities’ delay observed in this study is relatively short
though it still needs to be reduced further. It indicates that the TBMUs are doing
an encouraging job in the region.
Over all, considering the high magnitude of delay in the region, the above-
mentioned factors should be an area of focus for the regional health bureau to
start acting on them so as to lower down the current unacceptable long duration
of pretreatment period and reduce transmission of TB in the community.
Therefore, the following recommendations are made based on our findings.
Recommendations
1. There should be an access for a simple and rapid diagnostic test for TB at
the lowest health care facilities (health posts/clinic) so as to reduce the
health systems’ delay.
2. A mechanism has to be created to work closely with all non-formal health
providers including drug retail outlets, traditional healers, herbalists and
religious healers in the region on how to identify TB suspects and instant
referral. This could be done through workshops, seminars, conferences
and trainings organized by the RTLCP.
3. It is important that health workers working at the peripheral health care
facilities be more alert to the possibility of pulmonary TB in patients with
respiratory symptoms. So that it can be diagnosed early and treated
promptly to reduce patient morbidity as well as to limit its spread in the
community. Efforts should be made to improve the diagnostic skill and the
awareness of TB among all health workers particularly, nurses and health
assistants since most patients first seek treatment from them. Education
on the clinical identification of suspects, public health aspects of TB and
the importance of referral should be intensified at undergraduate level in
the nursing schools. There should also be continuing medical education
about TB in the form of lectures, conferences or seminars to other health
workers such as health officers and medical doctors working at the
diagnosing facility level to maintain a high index of suspicion for TB and
perform appropriate diagnostic tests. These all might be organized by the
RTLCP, NGOS and professional organizations working in the region
4. Well-organized and integrated information, education and communication
(IEC) program has to be put in place to raise the awareness of the
population in general on the symptoms and treatment of TB and facilitate
prompt utilization of the available health service. Using the available
media such as the regions radio and newspaper for the dissemination of
health information to the general population should be given due
emphasis.
5. Increased knowledge of patients’ health seeking behavior and their self-
perception of disease is useful for health workers and should have
implications for health education messages
6. In this study, far distance between home and health care facility has
affected the early initiation of treatment. Therefore, efforts should include
improving easy access of diagnostic facilities to the population by further
decentralizing the TB diagnostic and treatment services to the periphery
are necessary. Besides this, there should be a mechanism to collect
sputum samples from the remote areas. In this regard, incorporating the
community health workers might be important.
7. The regional TB control program has to device a system for the private
sector to effectively participate in TB control activities. This could help in
reducing health systems’ delay. Moreover, it is important to give training to
the private medical providers about the clinical and public health aspects
of TB.
8. Regular and intensive supervision including the government and the
private medical providers should be strengthened to assure the quality of
care given to TB patients. The supervision should also include all drug
retail outlets aiming at controlling the availability of anti-TB drugs in the
private pharmacies.
9. In this study, we observed that due to the wrong perception of the causes
and symptoms of TB, patients were treating themselves for a considerable
period of time prior to presenting to medical providers. Besides this, the
health service coverage in the region is low. Therefore, with only using
passive case finding, it might be difficult to reach as many pulmonary TB
patients as possible. Therefore, we believe that it is important to
incorporate active case finding, like contact tracing in the current TB
control system in the region.
4.12 Further research implications1 Further study should be conducted to see the relation between
longer pre-treatment delays and its effect on treatment outcome on
the already studied patients.
2 The magnitude of individuals having suspected symptoms of TB but
did not seek treatment should be explored to better understand the
impact of diagnostic delay in the case finding activity in the region.
3 As described in our introduction section, the report of the regional
state health bureau has revealed that the case holding activity in
the region has currently encountered challenges in its
implementation. Some of the challenges include poor recording and
reporting, poor patient follow-up, high staff turn-over and in general
poor management of DOTS. Therefore, looking into the quality of
TB control in the region might help to understand the underlying
problems and improve the TB control program in the region.
4 The role of private medical providers in TB diagnosis in the region
needs to be assessed. This will help to generate information that
can be used for designing better cooperation between the
government and the private sector in TB control in the region.
5. In this study we have observed that some of the respondents were
able to access anti-TB drugs from various drug retail outlets. We
know that this is against the guideline of the NTCPs as it can lead
to irrational use of anti-TB drugs. Therefore, we think it is wise to
conduct a baseline survey for assessing anti-TB drug resistance in
the region.
6. Information regarding TB and HIV co-infection is lacking in the
region. In the present study we found 47% of the TB patients
associating TB with HIV. Therefore, looking into the magnitude of
HIV/TB co-infection in the region might be useful in generating
information that can be used to plan a coordinated intervention
strategies for both diseases.
REFERENCES
1. Harlan R. Crops and man. First edition. Wisconsin (USA): American society for
agronomy (Madison); 1975. p. 69-72.
2. Central Statistics Authority (CSA). Demographic health survey in Ethiopia. Addis
Ababa 2000. p.1-30.
3. Ethiopian Ministry of Health. Health and health related indicators. Planning and
programming department: Addis Ababa: Ministry of health; 2002. p. 3-26.
4. Ethiopian Ministry of Health. Health policy of the transitional government of
Ethiopia. Addis Ababa: Ministry of Health; 1993. p. 21-37
5. PHRO. Policy and Human Resource Development Project Office (PHRO). Ethiopian
social sector studies, health sector review: synthesis and summary. Government
of Ethiopia and World Bank collaboration, Addis Ababa, Ethiopia, 1998.
6. World bank. Development report: Investing in health. New York, Oxford University
press, 1993.
7. Bishaw M. Promoting traditional medicine in Ethiopia. A brief historical review of
government policy. Soc Sci med; 1991. 32(2): 193-200
8. Yong A. Internalizing and externalizing medical belief systems: An Ethiopian
example. Soc Sci Med 1976; 10: 156-170.
9. Meseret S. The role of health professionals in the development of traditional
medicine in Ethiopia. In: Proceedings of the Work shop on Development and
utilization of Herbal Remedies in Ethiopia. Nazareth, Ethiopia 1996; 15-18
10. Ethiopian Ministry of Health. Manual for tuberculosis and leprosy control program.
3rd ed. Addis Ababa: Ministry of Health; 2002. p.1-20.
11. Amhara Regional State Health Bureau. Annual health service report. Bahirdar:
Amhara regional state health bureau; 2003. p.1-15
12. Amhara Regional State Health Bureau. Annual health service report. Bahirdar:
Amhara regional state health bureau; 2002. p.1-30
13. Amhara Regional State Health Bureau. Annual tuberculosis and leprosy control
program report. Bahirdar: Amhara Regional State Health Bureau; 2003.
14. World Health Organization. Global tuberculosis report for 2003. Geneva, Switzerland:
WHO (WHO/HTM/TB/03. 331); 2003.
15. World Health Organization. TB as a public health problem. Department of
communicable disease. (2screens) Available at:
http://www.whosea.org/tb/publichealth.htm. Accessed. January 21/2004.
16. Lawn S D, Afful B, Acheampong J W. Pulmonary tuberculosis: diagnostic delay in
Ghanaian adults. Int J Tuberc Lung Dis 1998; 2(8): 635-640.
17. Pirkis JE, Speed BR, Yung AP, Dunt DR, Mac Intyre CR, Plant AJ. Time to
initiation of anti-tuberculosis treatment. INT J Tuberc Lung Dis 1996 Oct; 77(5):
401-6.
18. Steen T. W, Mazonde GN. Pulmonary tuberculosis in Kweneng district, Botswana:
delays in diagnosis in 212 smear-positive patients. Int J Tuberc Lung Dis 1998;
2(8): 627-634.
19. Lienhardt C, Rowley J, Manneh K, Lahai G, Needham D, Milligan P, Mc Adam KP.
Factors affecting time delay to treatment in a tuberculosis control program in a
sub-Saharan African country: the experience of the Gambia. INT J Tuberc Lung
Dis 2001; 5(3): 233-9.
20. World Health Organization. Executive summary of meeting report, “Diagnosis of
tuberculosis: Countdown to New Tools.” 29-30 June 2000; (3 screens)
9. How many times have you visited a government medical provider for your symptoms before it was confirmed to be TB?
1. Once□ 2.Twice □ 3.Three times □ 4. Four times □ 5. Five times
□ 6. More than five□.
9.1Were the above visits with the same or different medical providers?
1. Same□ 2.Different
□9.2 Did the doctor ordered investigations for you at that time of illness? 1. Yes □
2.No □9.3 If yes, which of the following investigations were done?
Sputum examination 1. Yes □ 2.NO □
Chest x- ray 1.Yes □ 2. NO □
3. I do not remember□
10. How many times have you visited a private medical provider for your symptoms before it was confirmed to be TB?
1. Once□ 2.Twice □ 3. Three times □ 4. Four times □ 5.five times □ 6. More than five□.
10.1Were the above visits with the same or different private medical providers?
1. Same □ 2. Different □
10.2 Did the physician at the private medical provider ordered investigations for you at
that time of illness? 1.Yes □ 2. NO □
10.3 If yes, which of the following investigations were done?
Sputum examination 1.Yes □ 2. NO □
Chest x-ray 1.Yes □ 2. NO □
I do not remember□
11. Where did it become for the first time clear that the disease is TB?
1. TBMU □
2. Private medical provider □
D. If the diagnosis of TB was made at the private medical providers
12. What did the doctor/ the health worker at the private medical provider do when he/she confirmed that your illness was TB?
1. He/she referred me to the TBMU with slides □ 2. I was referred with out slides □ 3. I was given a prescription and sent to a pharmacy to buy anti TB drugs □
12.1 If given prescription, Did you get the drugs in the private pharmacy?
1. Yes □ 2. No □
12.2 If yes, did you purchase? 1. Yes □ 2.No □
If no, why not?
1. It was expensive □
2. I thought it was fake □3. Other Please describe ______________________
13. How long did it take from the time you were referred by the private medical provider till you first reported to the TBMU?_______days/weeks.
12.4 When you reached at the government (TBMU) with your referral, what did they do?
1. Re examined me □
2. Requested AFB □3. They accepted my slides and started me on treatment □4. Other, please describe________________________________
14. How long did it take from the time you first reported to the TBMU till you first started anti-TB drugs? ____________days/weeks
15. How long did it take from the time you were referred by a medical provider till you first started taking the anti TB drugs?_________________________days/weeks.
E. Diagnosis made at theTBMU
16. Date of first visit to the TBMU? __________________________________________
17. How did you decide to visit the TBMU? __________________________________
1. Referred by HP/clinic □ Date Referred________________
2. Self-Referred □ Date referred_________________
3. Referred by private □ Date referred_________________
24.2 Date of registration for treatment (from district registry book)Checked_____________________________________________________________
25. How long did it take from onset of the present illness till you first started anti TB chemotherapy? ____________________________ (days, weeks, month)
26. How much money did you pay for all the consultations & medications from onset of
cough till the diagnosis of TB? _____________Birr.
E Knowledge of TB:
27. Have you heard, known something about pulmonary TB? For example, TB causes
chronic cough? Haemoptysis? 1. Yes□ 2. No □
27.1. If yes, where has the information come from?
28. If TB is treated, can it be cured? 1. Yes□ 2. No□ 3.I do not know □
29. What do you think are causes of TB?Possible causes No yes I do not knowWitchcraftPovertyBacillihard workSexual overindulgenceMalnutritionUnventilated homeLiving together with untreated TB patientHIVother causes
31. Do you know any danger if a TB patient is not treated? 1.Yes □ 2.
No □
31.1 If yes, what is it?For the patient, ____________________________________________For the people around, ________________________________________
32. Do you know that the drugs are available free? 1. Yes □ 2.No□ 3. I don't
know □
33. How long is TB treated? 1. 1-year □ 6-8 months 2.□ I do not
know □
3. Other, please describe__________________
F. Stigma:
34. Do you feel TB is a social stigma? 1. Yes □ 2. No□
35. Before you came to this health facility, was there any fear in your mind that you
would be tested for HIV? 1. Yes □ 2.
No□
36. Do you think people will avoid your company because you are a TB patient?
1. Yes □ 2. No□
37. In your opinion going to the health center for TB test can make other people think that you have HIV/AIDS?
1. Yes□ 2.No□ 3. I do not know
□
38. Does TB has an association with HIV?
1. Yes □ 2.No □ 3. I do not know
□
39. Do you fear not to enter others social circle in fear that they will not accept you?