Prevalence and determinants of antibiotics self-medication among indigenous people of Bangladesh: a cross-sectional study

Abstract

Objectives

Self-medication with antibiotics (SMA) contributes significantly to the emergence of antimicrobial resistance (AMR), especially in low-income countries including Bangladesh. This study aimed to generate evidence on the self-reported prevalence of antibiotic self-medication and its determinants among indigenous people residing in Bangladesh’s Chittagong Hill Tracts (CHT) districts.

Design

This study used a cross-sectional design with data collected through a survey using a semi-structured questionnaire.

Setting

This study was conducted from late January to early July 2021; among different indigenous group populations aged 18 years or more olders residing in the three districts of CHT.

Participants

A total of 1336 indigenous people residing in Bangladesh’s CHT districts were included.

Primary outcome and explanatory variables

The primary outcome measure was SMA while explanatory variables were socio-demographic characteristics, health status of participants, and knowledge of antibiotics usage and its side effects.

Results

Among the study participants, more males (60.54%) than females (51.57%) reported using antibiotics. The SMA rate was high among individuals with education levels below secondary (over 50%) and those in the low-income group (55.19%). The most common diseases reported were cough, cold and fever, with azithromycin being the most frequently used antibiotic. Levels of education, family income, having a chronic illness and place of residence were found to be the significant predictors of having good knowledge of antibiotic use as found in the ordered logit model. Findings from a logistic regression model revealed that men had 1.6 times higher odds (adjusted OR (AOR) 1.57; 95% CI 1.12 to 2.19) of SMA than women. Participants with ≥US$893 per month family income had lowest odds (AOR 0.14; 95% CI 0.03 to 0.64) of SMA than those who earned <US$417. Participants living in Rangamati districts had a lower risk of SMA (…) than those in Bandarban district. rate of SMA (AOR 0.52; 95% CI 0.30 to 0.90) than those in Bandarban district.

Conclusion

Male gender, family income, place of residence and knowledge of antibiotics were the significant predictors of antibiotic self-medication. Hence, it is important to streamline awareness-raising campaigns at the community level to mitigate the practice of SMA in indigenous people and ultimately address the devastating effects of Antimicrobial resistance (AMR) in Bangladesh.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study is the first qualitative study of its kind to assess antibiotic self-medication and its determinants among indigenous people in Bangladesh.

• This study included a robust sample with variables and indigenous people from three different districts.

• This study will guide the development of national policies and guidelines for addressing the self-medication of antibiotics.

• The study’s main weakness is that it is cross-sectional, meaning that only associations can be inferred and causality cannot be determined.

Introduction

While the discovery of antibiotics has markedly influenced treatment methods against infectious diseases, unfortunately, most available antibiotics are gradually becoming ineffective at a higher rate.¹ Antimicrobial resistance (AMR) is presently a significant concern of the World Health Organization (WHO),² accounting for thousands of deaths annualy. In Europe, the number of deaths from AMR is 33 000 per year,³ while in the USA, AMR causes as many as 35 900 deaths per year.⁴ The fatality is even worse in low- and middle- income countries (LMICs).^{5 6}

The emergence of AMR occurs due to evolutionary selective pressure exerted on the bacterial population due to the irrational use of antibiotics, causing the bacteria to circumvent the drug action and multiply.⁷ Treatment of highly prevalent diseases such as pneumonia, cholera, tuberculosis, and malaria has become difficult due to the emergence of AMR.⁸

Although the mechanism of AMR develops in bacteria naturally, human activities can accelerate the emergence and spread of the process. Previous studies have identified both high consumption of antibiotics and irrational consumption of the drug as contributing factors to AMR.⁹ A study has reported that approximately 50% of human and animal antibiotic usage is unreasonable.¹⁰ Often, the irrational use of antibiotics comes from several socio-demographic determinants. Recent reports show that people in developed countries with high education rates often engage in inappropriate antibiotic use through self-medication with antibiotics (SMA).^{11 12} Surprisingly, students from health-related programs in Malaysia showed ignorance regarding antibiotic usage.⁹ In LMICs, people have the same difficulties regardless of their level of education or economic standing. Unfortunately, there is widespread ignorance among the population regarding the distinctions between antibiotics and other types of treatment.^13–15 Because of their lack of knowledge, many believe antibiotics may treat any ailment, even viral diseases.¹⁶ In the USA, 29% of annual antibiotic prescriptions (~11.4 million) are unnecessarily prescribed for acute respiratory tract infections only.¹⁷ Sometimes, patients ask physicians to prescribe them antibiotics adamantly.¹⁸

Apart from education, age is a significant variable in terms of self-medication. Multiple studies have reported that cases of SMA are more prevalent among adults aged 18 to 45 years and older.^{19 20} Most studies, with a few notable exceptions, have found that SMA is more prevalent in females than males.^19–23 Furthermore, there is a correlation between the monthly family income and the tendency to engage in SMA.^{24 25}

Antibiotics are available without a prescription in many low- and middle-income countries, leading to the widespread occurrence of Self-Medication with Antibiotics (SMA).^{26 27} In Bangladesh, the SMA prevalence among the general population was reported to be 26.69%.²⁸ SMA was shown to be relatively frequent among the people who lived in rural parts of Dhaka, Chittagong and Rajshahi, according to cross-sectional studies that were conducted among participants who lived in those locations.^{29 30} An investigation in Dhaka revealed that 75% of the people polled used over-the-counter drugs, of which 16.90% were antibiotics.³¹ It is necessary to address the situation in low-income nations, including Bangladesh, to enhance the overall health impact, given that majority of the world’s population resides in these countries.³²

Bangladesh is a densely populated country where 1.8% of the total population comprises indigenous people living in the rural areas of Chittagong, Mymensingh, Sylhet and Rajshahi.³³ These ethnic groups have long maintained isolation from the general population, putting them at a disadvantage in obtaining healthcare facilities. As a consequence, their overall health status remains poorer compared with non-indigenous people.^34–36 There is not sufficient information regarding antibiotics usage, including SMA, among this community of more than 2 million people. It is possible that the indigenous population’s limited access to resources resulting from social, cultural, geographical, and financial abilities contributes to their increasing use of SMA.³⁴

This study was conducted among the indigenous people of Chittagong to generate evidence on the self-reported prevalence and determinants of antibiotic self-medication and its determinants. Age, gender, education level, occupation, income, and ethnicity are various socio-demographic criteria that have been taken into consideration to examine the overall background of SMA. The pattern of antibiotic usage, the reasons for SMA, the healthcare-seeking practices, as well as knowledge and attitudes regarding antibiotic usage have been investigated in this study. The findings of this study will help us address the knowledge gaps in the usage of non-prescribed antibiotics, pharmacy regulations, availability of antibiotics, knowledge of antibiotic use and awareness of antibiotic misuse in order to recommend necessary intervention and policy reform.

Methods

Participants

A cross-sectional study was conducted between January 2021 and July 2021 among indigenous people residing in Rangamati, Bandarban, and Khagrachari, three hill districts in the CHT in Bangladesh. The indigenous population was characterised per the description provided by the International Work Group for Indigenous Affairs (IWGIA).^{37 38} Within the study, the indigenous population was delineated into prominent ethnic groups, including Chakma, Marma, Tripura, Bawm, Tanchangya, Chak, Rakhaine, Saotal, and Mro. The categorisation of an individual’s ethnicity ensued through a systematic examination, integrating parameters such as self-identification, observational data, surname analysis, linguistic attributes, cultural practices, and geographical origin. Both men and women aged 18 years and older living in the selected study sites were included in the study. The sample size (n=383) for this study was determined using the formula n=z² p(1−p)/d²,³⁹ taking into account a significance level of 0.05, a confidence level of 95%, and a prevalence rate of 26.69% obtained from a prior study in Bangladesh, a 5% margin of error, a 20% non-response rate, and finally multiplied by a design effect 2 due to the cluster sampling technique.⁴⁰ All the respondents gave voluntary consent and agreed to participate in the study.

Sampling

This study used a two-stage cluster sampling approach in the districts of Bandarban, Rangamati and Khagrachari. Six out of seven upazilas (sub-districts) in Bandarban, nine out of ten in Rangamati and nine in Khagrachari were included in the study. Two sub-districts, Rowangchhari in Bandarban and Baghaichari in Rangamati, were excluded due to COVID-19 lockdown and security concerns. The samples were selected in proportion to the indigenous population based on the 2011 census: 20.3% from Bandarban, 37.5% from Khagrachari and 42.12% from Rangamati.^{41 42} In the first stage, the sampling units were the sub-districts, and a probability proportionate to size method was employed. This technique ensured that sub-districts with more households had a higher chance of being selected. For the second stage, participants (either the head of the household or another family member in the absence of the head) from the chosen sub-districts were identified using systematic random sampling. The data were collected through face-to-face interviews using a semi-structured questionnaire consisting of 60 questions arranged in four separate sections. The questionnaire consisted of four sections; section 1 assessed the demographic details and epidemiological characteristics including comorbidities, lifestyle and reasons for using antibiotics. Section 2 covers the pattern of antibiotic usage. Section 3 takes into account healthcare-seeking facilities. They were asked about hygiene, the frequency of respondents visiting physicians’ offices and the distance from the closest pharmacy. Section 4 assessed the knowledge and attitude of indigenous people toward antibiotic usage.

This study has recognized the self-medication of antibiotics (SMA) as an irrational use of antibiotics.⁴³ The WHO delineates self-medication as ‘the use of drugs to treat self-diagnosed disorders or symptoms, or the intermittent or continued use of a prescribed drug for chronic or recurrent disease or symptoms’.^{21 44} Survey respondents were asked whether they had consumed medication during the past year without consulting a doctor or any healthcare professional, which was identified as the primary outcome variable of this study. Interviews were conducted by research assistants after they had the necessary training on data collection and questionnaires. We recruited seven local research assistants fluent in the indigenous tribes’ languages and provided thorough training before data collection. The question was initially written in English and later translated into Bengali. The questionnaire was evaluated and reviewed by a panel of experts, including doctors, pharmacists, epidemiologists, and medical specialists. All data was initially recorded manually on paper and then entered into Google Forms for export and storage in Microsoft Excel 2013.

Explanatory variables

Explanatory variables comprised of administrative districts (Rangamati, Bandarban, and Khagrachari), age groups (18–29, 30–39, 40–49, 50–59, and 60 years and above), gender (male and female), ethnicity (Chakma, Marma, Tripura, Bawm, Tanchangya, Chak, Rakhain, Saotal, and Mro), educational levels (illiterate, primary, secondary, higher secondary, and graduate), occupations (agricultural work, service, housewife, student, business, day labour, handloom, and unemployed), family income in US dollars ($US) (<US$417, US$417–893, >US$893). Additional variables considered were the prevalence of common diseases (cough, cold and fever, headache, joint pain, diarrhoea and food poisoning, dental carries and toothache, irritable bowel syndrome, typhoid, malaria, jaundice, roundworm/tapeworm, sinusitis, asthma, other respiratory diseases, acne, skin allergies, and none) and types of medication concerning self-medication associated with disease prevalence.

Patient and public involvement

No patient was involved.

Statistical analysis

A descriptive analysis was performed to determine the distribution of socio-demographic variables. χ² tests with a significance threshold of 5% were used to compare the prevalence of self-medication across the variables. An ordinal logistic regression model was performed to determine the predictors of the knowledge level of antibiotic self-medication as a dependent variable. Knowledge level was categorized as no knowledge, poor knowledge, fair knowledge, and good knowledge. A binary logistic regression analysis was performed with ever-reported self-medication of antibiotics as a dependent variable, and independent variables included socio-demographic characteristics including age, gender, education, marital status, occupation, family income, and home district. Other independent variables included in the final model are having suffered from any chronic illness and knowledge about antibiotics. The variables having a p value of <0.25 in the univariate logistic regression analysis were included in a multivariable model to estimate the adjusted OR. Both unadjusted and adjusted OR are presented with 95% CIs. In the final model, a p-value of 0.05 or below was considered as statistically significant. SPSS V.25.0 was used for statistical analysis. GraphPad Prism V.9.0 was used to represent the findings graphically.

Result

Socio-demographic characteristics

A total of 1537 individuals took part in the study, of which 749 respondents (48.7%) used antibiotics, 587 respondents (38.1%) never used antibiotics, and the remaining (13.2%) were unsure if they used antibiotics (online supplemental figure 1). However, information on antibiotic use was available for 1336 of the 1537 participants. Among the 1336 individuals from distinct indigenous communities who participated in the study conducted in three hill districts of Bangladesh’s Chittagong division and reported use of antibiotics, were as follows: Rangamati (42.1%), Khagrachari (31.6%) and Bandarban (26.3%), with participation from Chakma (40.2%) and Marma (37.2%) followed by Tripura (9.4%), Tanchangya (6.3%), Bawm (5.5%) and other communities (1.4%) (table 1). The majority of the participants (46.3%) in this study were young (18–29 years), and the prevalence of using antibiotics was comparatively higher in this age group (table 1). The ratio of the male (50.1%) and female (49.9%) populations in this study was nearly similar among the study participants, and the male population (60.5%) reported using antibiotics more than females (51.57%). More than half of the participants had higher secondary or higher education, with major occupations including students (32.6%) and salaried job holders (23.1%). Diabetes and hypertension stand out as the most prevalent comorbidities within the SMA population (online supplemental figure 2).

Table 1.

Socio-demographic characteristics of study participants by antibiotic use

	Total (N=1336) N (%)	Did not use antibiotics (N=587) N (%)	Used antibiotics (N=749) N (%)
Age group (years)
18–29	619 (46.30)	241 (38.93)	378 (61.07)
30–39	255 (19.10)	131 (51.37)	124 (48.63)
40–49	207 (15.50)	100 (48.31)	107 (51.69)
50–59	149 (11.20)	69 (46.31)	80 (53.69)
60 and above	106 (7.90)	46 (43.40)	60 (56.60)
Gender
Female	667 (49.90)	323 (48.43)	344 (51.57)
Male	669 (50.10)	264 (39.46)	405 (60.54)
Education level
Illiterate	181 (13.60)	116 (64.09)	65 (35.91)
Primary school	222 (16.60)	114 (51.35)	108 (48.65)
Secondary school	221 (16.50)	98 (44.34)	123 (55.66)
Higher Secondary	319 (23.90)	132 (41.38)	187 (58.62)
Bachelor and above	393 (29.40)	127 (32.32)	266 (67.68)
Marital status
Married	744 (55.70)	354 (47.58)	390 (52.42)
Never married	547 (40.90)	207 (37.84)	340 (62.16)
Other	45 (3.40)	26 (57.78)	19 (42.22)
Occupation
Agriculture	255 (19.10)	145 (56.86)	110 (43.14)
Service	308 (23.10)	116 (37.66)	192 (62.34)
Other	68 (5.10)	32 (47.06)	36 (52.94)
Student	436 (32.60)	159 (36.47)	277 (63.53)
Housewife	219 (16.40)	120 (54.79)	99 (45.21)
Unemployed	50 (3.70)	15 (30.00)	35 (70.00)
Family income
<US$417	1150 (86.08)	530 (46.09)	620 (53.91)
US$417 to <US$893	161 (12.05)	52 (32.30)	109 (67.70)
≥US$893	25 (1.87)	5 (20.00)	20 (80.00)
Ethnicity
Chakma	537 (40.20)	207 (38.55)	330 (61.45)
Marma	497 (37.20)	213 (42.86)	284 (57.14)
Tripura	126 (9.40)	70 (55.56)	56 (44.44)
Tanchangya	84 (6.30)	41 (48.81)	43 (51.19)
Bawm	74 (5.50)	49 (66.22)	25 (33.78)
Other	18 (1.40)	7 (38.89)	11 (61.11)
Home district
Bandarban	351 (26.30)	240 (68.38)	111 (31.62)
Khagrachari	422 (31.60)	145 (34.36)	277 (65.64)
Rangamati	563 (42.10)	202 (35.88)	361 (64.12)

Pattern of self-medication of antibiotics usage

Only 7 of the 749 participants were excluded from the study because it was unclear whether they used antibiotics that had been prescribed by a doctor. Approximately 382 respondents (51.4% of 742) who used antibiotics did not have prescriptions. Self-medication for antibiotic usage was significantly different by gender, education level, occupation, and family income (table 2). The SMA rate was higher than 50% in the illiterate (58.46%), primary school (63.55%) and secondary school (54.55%) groups. The low-income group (55.19%) also had a higher SMA rate.

Table 2.

Comparison of respondents by antibiotic usage with and without prescription

	Total (N=742) N (%)	Self-medication (N=382) N (%)	Prescription (N=360) N (%)	P value
Age group (years)				>0.05
18–29	375 (50.54)	204 (54.40)	171 (45.60)
30–39	123 (16.58)	58 (47.15)	65 (52.85)
40–49	107 (14.42)	55 (51.40)	52 (48.60)
50–59	77 (10.38)	34 (44.16)	43 (55.84)
60 and above	60 (8.09)	31 (51.67)	29 (48.33)
Gender				<0.01
Female	340 (45.82)	153 (45.00)	187 (55.00)
Male	402 (54.18)	229 (56.97)	173 (43.03)
Education level				<0.05
Illiterate	65 (8.76)	38 (58.46)	27 (41.54)
Primary school	107 (14.42)	68 (63.55)	39 (36.45)
Secondary school	121 (16.31)	66 (54.55)	55 (45.45)
Higher Secondary	184 (24.80)	91 (49.46)	93 (50.54)
Bachelor and above	265 (35.71)	119 (44.91)	146 (55.09)
Marital status				>0.05
Married	388 (52.29)	194 (50.00)	194 (50.00)
Never married	336 (45.28)	179 (53.27)	157 (46.73)
Other	18 (2.43)	9 (50.00)	9 (50.00)
Occupation				<0.01
Agriculture	110 (14.82)	70 (63.64)	40 (36.36)
Service	187 (25.20)	79 (42.25)	108 (57.75)
Other	36 (4.85)	23 (63.89)	13 (36.11)
Student	275 (37.06)	141 (51.27)	134 (48.73)
Housewife	99 (13.34)	47 (47.47)	52 (52.53)
Unemployed	35 (4.72)	22 (62.86)	13 (37.14)
Family income				<0.001
<US$417	616 (83.02)	340 (55.19)	276 (44.81)
US$417 to <US$893	106 (14.29)	40 (37.74)	66 (62.26)
≥US$893	20 (2.70)	2 (10.00)	18 (90.00)
Chronic condition				<0.01
No	296 (39.89)	171 (57.77)	125 (42.23)
Yes	446 (60.11)	211 (47.31)	235 (52.69)
Ethnicity				<0.05
Chakma	326 (43.94)	164 (50.31)	162 (49.69)
Marma	281 (37.87)	148 (52.67)	133 (47.33)
Tripura	56 (7.55)	19 (33.93)	37 (66.07)
Tanchangya	43 (5.80)	30 (69.77)	13 (30.23)
Bawm	25 (3.37)	15 (60.00)	10 (40.00)
Other	11 (1.48)	6 (54.55)	5 (45.45)
Home district				<0.05
Bandarban	111 (14.96)	71 (63.96)	40 (36.04)
Khagrachari	273 (36.79)	133 (48.72)	140 (51.28)
Rangamati	358 (48.25)	178 (49.72)	180 (50.28)

*P values were analysed using χ² tests, and p<0.05 was considered statistically significant.

Knowledge of the use of antibiotics

The results reported that education level, occupation, family income, chronic morbidity, and home district significantly predict the knowledge level (no, poor, fair, or good) (p≤0.001) (table 3). The participants having highest level of education had the best likelihood of having good knowledge of antibiotic use compared to the group who were illiterate (p≤0.05). Among the various occupation groups, job holders or students had good knowledge about antibiotic use than those who work in agriculture (p≤0.01). Compared with the individuals who earned <US$417 had a higher likelihood of having good knowledge of antibiotic use than those who earned ≥US$893 (p<0.05). In addition, patients who suffered from chronic ailments were more likely to deeper comprehension of the use of antibiotics (p<0.001). Residents from Rangamati districts had the highest likelihood of knowing the use of antibiotics compared with those who live in Bandarban (p<0.001).

Table 3.

Ordered logit estimates for knowledge level on antibiotic use

	Not at all	Poor	Fair	Good	Knowledge level
	(N=478)	(N=577)	(N=394)	(N=88)	(N=1537)
	N (%)	N (%)	N (%)	N (%)	(Ordered logit)
Age group (years)
18–29	145 (20.00)	294 (40.55)	236 (32.55)	50 (6.90)	Reference
30–39	106 (36.55)	108 (37.24)	60 (20.69)	16 (5.52)	0.9171
40–49	90 (38.63)	98 (42.06)	39 (16.74)	6 (2.58)	0.8268
50–59	68 (40.96)	49 (29.52)	38 (22.89)	11 (6.63)	0.9732
60 and above	69 (56.10)	28 (22.76)	21 (17.07)	5 (4.07)	0.9167
Gender
Female	266 (35.23)	254 (33.64)	196 (25.96)	39 (5.17)	Reference
Male	212 (27.11)	323 (41.30)	198 (25.32)	49 (6.27)	0.8899
Education level
Illiterate	147 (73.50)	45 (22.50)	8 (4.00)	0 (0.00)	Reference
Primary school	145 (57.77)	77 (30.68)	25 (9.96)	4 (1.59)	2.1251***
Secondary school	90 (35.43)	116 (45.67)	38 (14.96)	10 (3.94)	4.1917***
Higher Secondary	64 (16.71)	165 (43.08)	131 (34.20)	23 (6.01)	8.6442***
Bachelor and above	32 (7.13)	174 (38.75)	192 (42.76)	51 (11.36)	15.8541***
Marital status
Married	339 (40.36)	295 (35.12)	169 (20.12)	37 (4.40)	Reference
Never married	110 (17.05)	268 (41.55)	218 (33.80)	49 (7.60)	0.9076
Other	29 (55.77)	14 (26.92)	7 (13.46)	2 (3.85)	0.7553
Occupation
Agriculture	170 (60.93)	86 (30.82)	17 (6.09)	6 (2.15)	Reference
Service	65 (19.35)	125 (37.20)	121 (36.01)	25 (7.44)	1.8889**
Other	31 (38.27)	36 (44.44)	11 (13.58)	3 (3.70)	1.2036
Student	66 (12.52)	219 (41.56)	192 (36.43)	50 (9.49)	1.6798***
Housewife	123 (49.80)	85 (34.41)	36 (14.57)	3 (1.21)	0.9769
Unemployed	23 (34.33)	26 (38.81)	17 (25.37)	1 (1.49)	0.8305
Family income
<US$417	456 (34.29)	519 (39.02)	294 (22.11)	61 (4.59)	Reference
US$417 to <US$893	18 (10.11)	48 (26.97)	95 (53.37)	17 (9.55)	1.9034***
≥US$893	4 (13.79)	10 (34.48)	5 (17.24)	10 (34.48)	2.2534**
Chronic condition
No	284 (35.90)	296 (37.42)	169 (21.37)	42 (5.31)	Reference
Yes	194 (26.01)	281 (37.67)	225 (30.16)	46 (6.17)	1.4551***
Ethnicity
Chakma	156 (24.19)	218 (33.80)	215 (33.33)	56 (8.68)	Reference
Marma	197 (34.62)	230 (40.42)	123 (21.62)	19 (3.34)	0.9112
Tripura	44 (32.84)	42 (31.34)	40 (29.85)	8 (5.97)	1.3464
Tanchangya	35 (36.08)	48 (49.48)	10 (10.31)	4 (4.12)	0.9263
Bawm	34 (45.95)	36 (48.65)	4 (5.41)	0 (0.00)	1.1154
Other	12 (66.67)	3 (16.67)	2 (11.11)	1 (5.56)	0.5065
Home district
Bandarban	168 (46.15)	155 (42.58)	30 (8.24)	11 (3.02)	Reference
Khagrachari	164 (33.40)	137 (27.90)	148 (30.14)	42 (8.55)	1.7414***
Rangamati	146 (21.41)	285 (41.79)	216 (31.67)	35 (5.13)	2.5839***

*p<0.1; **p<0.05; ***p<0.001.

Reasons and antibiotics usage among the participants

The top three reasons for SMA were previous experience (36.5%), the urgency of the problem (22.6%) and economic constraints (16.9%) in this study (figure 1A). Around 43% of the individuals took antibiotics as advised by the pharmacies, while 38% followed written prescriptions for using antibiotics. For taking antibiotics without a prescription, participants in the study population demonstrated that using previous experience was the most common reason for SMA practice (36.54%) (figure 1A). Long distances between home and facilities (22.66%), high consultation fees, and the urgency to treat symptoms were reasons for using antibiotics without a valid prescription. Other sources include advice without a prescription from primary care physicians, previous prescriptions, experience and more. Cough, cold, and fever (70.1%); headache (13.8%); dysentery, diarrhoea and food poisoning (11.8%) were the top three diseases behind antibiotic use (figure 1B). The most commonly prescribed antibiotics for the above-mentioned disease behind the use of antibiotics were azithromycin (27.0%), amoxicillin (20.0%), metronidazole (18.0%), followed by ciprofloxacin, cefixime, tetracycline, etc (figure 1C).

Figure 1.

Characteristics of self-prescribed antibiotic usage and common disease that antibiotic was used for. (A) The reason behind the use of self-prescribed antibiotics. (B) Common diseases behind antibiotics. (C) Mostly self-prescribed antibiotics where the X-axis represents the antibiotics name and the Y-axis represents the percentage of antibiotics usages.

Determinants of antibiotic self-medication

The male gender was 1.6 times more likely to self-medicate with antibiotics compared with females (AOR 1.57; 95% CI 1.23 to 2.19). Participants with high income were significantly less likely to use self-medicating antibiotics (AOR 0.14; 95% CI 0.03 to 0.64). Those who knew the difference between antibiotics and other drugs were significantly less likely to self-medicate antibiotics (AOR 0.45; 95% CI 0.34 to 0.74). Indigenous people living in the Rangamati districts were less likely to self-medicate with antibiotics (AOR 0.52; 95% CI 0.30 to 0.90) compared with the residents from Bandarban (table 4). There was no statistically significant difference in antibiotic self-medication across ethnic groups among the indigenous population that participated in the study.

Table 4.

Multivariate analysis of factors associated with self-medication with antibiotics

	Unadjusted OR (95% CI)	P value	Adjusted OR (95% CI)	P value
Age group (years)
18–29 (ref.)
30–39	0.66 (0.35 to 1.24)	0.197	0.63 (0.33 to 1.20)	0.159
40–49	0.75 (0.386 to 1.48)	0.412	0.71 (0.36 to 1.41)	0.326
50–59	0.54 (0.26 to 1.12)	0.099	0.57 (0.27 to 1.21)	0.143
60 and above	0.69 (0.0307 to 1.54)	0.364	0.72 (0.31 to 1.64)	0.433
Gender
Female (ref.)
Male	1.60 (1.15 to 2.22)	0.005	1.57 (1.13 to 2.14)	0.008
Education level
Illiterate (ref.)
Primary school	1.27 (0.64 to 2.52)	0.485	1.32 (0.66 to 2.63)	0.429
Secondary school	0.86 (0.43 to 1.73)	0.670	0.98 (0.48 to 1.99)	0.948
Higher Secondary	0.67 (0.32 to 1.41)	0.293	0.86 (0.40 to 1.85)	0.701
Bachelor and above	0.57 (0.27 to 1.17)	0.126	0.76 (0.35 to 1.63)	0.482
Marital status
Married (ref.)
Never married	0.85 (0.46 to 1.56)	0.595	0.83 (0.44 to 1.55)	0.554
Other	1.29 (0.46 to 3.61)	0.624	1.22 (0.43 to 3.44)	0.704
Occupation
Agriculture (ref.)
Service	0.67 (0.37 to 1.22)	0.189	0.75 (0.41 to 1.36)	0.344
Other	1.37 (0.58 to 3.19)	0.470	1.31 (0.55 to 3.09)	0.545
Student	0.98 (0.50 to 1.90)	0.952	1.05 (0.535 to 2.06)	0.895
Housewife	0.78 (0.43 to 1.43)	0.424	0.84 (0.45 to 1.55)	0.579
Unemployed	1.33 (0.55 to 3.21)	0.528	1.36 (0.55 to 3.34)	0.500
Family income
<US$417 (ref.)
US$417 to <US$893	0.64 (0.40 to 1.02)	0.062	0.70 (0.44 to 1.12)	0.139
≥US$893	0.11 (0.02 to 0.51)	0.005	0.14 (0.03 to 0.64)	0.011
Ethnicity
Chakma (ref.)
Marma	0.89 (0.61 to 1.29)	0.533	0.97 (0.66 to 1.42)	0.863
Tripura	0.47 (0.25 to 0.91)	0.026	0.53 (0.27 to 1.02)	0.059
Tanchangya	1.31 (0.61 to 2.81)	0.491	1.44 (0.66 to 3.11)	0.355
Bawm	0.67 (0.24 to 1.84)	0.436	0.66 (0.24 to 1.82)	0.420
Other	0.88 (0.25 to 3.14)	0.842	0.97 (0.27 to 3.53)	0.968
Home district
Bandarban (ref.)
Khagrachari	0.53 (0.30 to 0.94)	0.029	0.57 (0.32 to 1.02)	0.057
Rangamati	0.47 (0.28 to 0.81)	0.006	0.52 (0.30 to 0.90)	0.020
Chronic condition
No (ref.)
Yes			0.73 (0.52 to 1.01)	0.060
Know about antibiotics
No (ref.)
Yes			0.74 (0.52 to 1.07)	0.111
Knows about antibiotics and other drugs
No (ref.)
Yes			0.50 (0.34 to 0.74)	0.001
Constant	3.62 (1.32 to 9.90)	0.012	4.33 (1.53 to 12.24)	0.006
No of observations	742		742
Pseudo r-squared	0.073		0.095

The dependent variable ‘self-prescription’ assumes a value of 1 if the respondent reported taking antibiotics that were self-prescribed and assumes a value of 0 if not.

Most (12.5%) of the self-prescribed antibiotic users had diabetes (online supplemental figure 3). Other chronic diseases reported were hypertension (8.0%), anxiety (7.0%), eye problems (6.5%), heart disease (5.0%), and respiratory disease (5.0%).

Discussion

This study showed that more than half of the antibiotic users from the indigenous population in the Chittagong Hill Tracts (CHT) practiced self-medication. This proportion was noticeably higher than other studies conducted in Dhaka and Rajshahi.^{40 45} The contrasting SMA prevalence within the country could be due to different target populations, cultural diversities, and socio-economic status. This deviation from intra-country variation highlights the need to monitor the legislation of drug prescription, delivery and consumption, even in the remote areas of Bangladesh to fight the baneful spread of AMR. These statistics of SMA are also higher than other countries—South-western India (18.0%), Nepal (26.2%), Turkey (19.0%), Ethiopia (22.7%), and Greece (20.0%).^{39 46–50}

Within the sample population, the socio-demographic factors that are associated with SMA were explored as part of the current study. The low-income group (<US$417) has a higher level of SMA practice. People in this poor economic condition rarely visit qualified doctors at a government health centre as it requires transport costs and time expenditure, leading to SMA.⁵¹ Additionally, owing to a high prevalence of poverty, many individuals within these communities are unable to purchase the complete course of antibiotics, driving an inclination towards non-prescription antibiotic use. The easy accessibility of certain over-the-counter antibiotics also contributes to SMA.³²

In our study, we found that many of the respondents consumed antibiotics previously prescribed by doctors for the treatment of a different disease. In some cases, a positive outcome in terms of recovery influenced them to continue using the same antibiotic to treat similar symptoms, even if it is a simple cough and cold (figure 1B). This fact is in concordance with non-indigenous Bangladeshi people.⁴⁰ Depending on past experience, it may not be a rational approach if the antibiotic was previously used to treat a different type or severity of infection, particularly in terms of the dosage and duration of treatment. To prevent unnecessary prescriptions of antibiotics and promote adherence to guidelines, physicians should restrict the issuance of excessive prescriptions, as many patients rely on prior prescriptions for information regarding medications.¹⁹

Previous studies have highlighted a higher prevalence of SMA in older people compared with the younger age cohort.⁵² The higher incidence of SMA in elders may be due to the accumulation of illness episodes and/or previous experiences with SMA, which are factors that contribute to the development of SMA.⁵² This study found that the 18–29 year-old cohort of indigenous people residing in the CHT were the largest antibiotic consumers. A similar study reported the same age group as the largest antibiotic consumer in the rural non-indigenous population of Bangladesh.⁵³ Conversely, non-indigenous young people in urban areas do not use antibiotics at that level.⁵⁴ This could be explained by the fact that adults who become independent are more likely to make self-decisions rather than relying on a physician’s decision regarding SMA.⁴⁰ Early health education about the proper use of antibiotics and personal hygiene can help the group develop lifelong healthy habits. By focusing on them, we also have a chance of reaching other family members to convey the message.⁴⁷

In line with past studies in Bangladesh, socio-demographic analysis of the sample population revealed that men consume more antibiotics than women do.^{40 55 56} However, these findings deviate from most of the results found in a similar study in Ethiopia, where the prevalence of SMA was 54.6% among female participants.⁴⁸ The difference in results also indicates that females are more cautious about health decisions than males. The gender disparity in SMA may stem from the fact that women, in general, in LMICs do not have equal access to healthcare compared with men.⁵⁷ The determination of whether males or females possess more knowledge and exhibit superior practices is heavily influenced by the specific situation and is contingent on several socio-demographic characteristics. Consequently, it is not possible to make generalisations about gender differences. Although the findings are not always consistent, the majority of studies did confirm the existence of gender disparities.⁵⁸

Education level is crucial in SMA; ^{19 59} studies in developing and developed countries also reported family income as a strong predictor of antibiotic knowledge.^60–62 Regardless of education, family income also regulates people’s access to information.⁶² However, people from both higher and lower education strata reported practicing self-prescribed antibiotics in LMICs.^63–65 Illiterate or low-educated people are rarely aware of health risk issues as well as SMA-associated threats. In contrast, highly educated people and health sector students hold false confidence, which drives them to SMA.²⁵ The present study depicts that most of the illiterate people (73.5%) were not aware of antibiotic use at all. A correlation between antibiotic knowledge and educational qualifications was reported in other countries.^{66 67} It is unlikely for less educated people to acquire antibiotic knowledge from newspapers, magazines, and websites. Efforts must be made to reach out to these people and disseminate antibiotic knowledge that will ultimately reduce the rate of AMR in CHT people. Antibiotic knowledge is also correlated with occupation. Job holders and students had better knowledge of antibiotics, as also reported in other studies.^68–70 The segment of the study population that is more connected to the media, the internet, and other social communication systems had better knowledge of antibiotic usage. On the contrary, farmers⁷¹ and housewives⁷² from CHT were either illiterate or poorly educated and demonstrated no understanding of the far-reaching consequences of SMA.

Patients with chronic diseases had better understanding of antibiotic usage (table 3). Antibiotics are administered for extended courses of treatment for patients with chronic disease. They are forced to use these medicines to a significant extent throughout this time period.⁷³ Because of this lengthy usage, their expertise regarding this topic has increased. Even family members of just one or two chronic patients were found to have a better understanding of antibiotics than those who do not have such patients in their family.^{74 75}

Survey respondents from Khagrachari and Rangamati exhibit better antibiotic knowledge than those from Bandarban. Such a difference might be a because of community clinic (CC) activities. The government of the People’s Republic of Bangladesh established CCs in 1998.⁷⁶ These CCs try to make people aware of the rational use of antibiotics.⁷⁶ The total number of CCs in Bandarban is less in Khagrachari or Rangamati (CCs in Bandarban, Khagrachari and Rangamati: 92, 108 and 118, respectively).⁷⁷

This study indicate that the study population voluntarily used antibiotics for cough, cold, or fever in more than 70% of cases. The use of antibiotics in such a pattern accelerated the development of AMR.⁹ This study revealed that the most commonly used antibiotics for self-medication are azithromycin (27.0%), amoxicillin (20.0%), metronidazole (18.0%), and ciprofloxacin (14.0%), which belong to the macrolides, penicillins, nitroimidazole, and fluoroquinolones classes of antibiotics, respectively.^{78 79} Alarmingly, most of them are broad-spectrum in nature, which poses the threat of developing a wider and more varied range of antibiotic-resistant organisms.⁸⁰

One more crucial and essential result that came out of this research was the identification of the reasons that contribute to SMA. These factors include a lack of easy access to healthcare, high consultation costs, extensive distances between consultation centres, and prescription-free medication sold in pharmacies. One of the most effective ways to address these issues is to increase the number of CCs that are staffed full-time by appropriately trained medical professionals. At the same time, it is imperative that pharmacies ban the practice of selling antibiotics to customers without a prescription. It may be helpful to reduce the rate of SMA by counselling people to help them understand that previous usage of antibiotics should not be the basis of SMA for future illnesses. Several questions, such as why self-medication is more prevalent in the Bandarban area and among the Tanchangya population, warrant further investigation to identify underlying issues at the root level. Moreover, it would be beneficial to undertake additional research to investigate the factors contributing to the prevalence of SMA among the individual indigenous populations under study separately. Additionally, several studies are required to obtain a complete picture of the misuse of antibiotics in this population. For example, what circumstances cause local medicine vendors to sell antibiotics without a prescription? Is it a lack of awareness, lax law enforcement, or simply taking advantage of underprivileged people? Is there an alternative method that could be more effective in conveying the message that the voluntary use of antibiotics can have potentially harmful consequences? Through investigation of these questions, it will be possible to gain a more in-depth understanding of the situation and find a solution to the problem of AMR. These, in turn, will assist in reducing the risk of the impending threat among the indigenous community of Bangladesh.

Limitations

This study has a few limitations that needs attention. This study followed a cross-sectional design that may introduce some level of recall bias on antibiotic usage and types of antibiotics. Thus, the study findings should be interpreted with caution. Furthermore, a few areas of Khagrachari and Bandarban were excluded from the study for security reasons, which may limit the findings being generalizable across the indigenous population.

However, this study was unique in terms of recruiting a sample size, highlighting antibiotic self-medication among indigenous people in three districts in Bangladesh. Study findings further highlighted that a high proportion of antibiotic self-medication needs immediate attention and further confirmation by larger studies. Nevertheless, the evidence revealed in this study will be useful for policy formulation and raising awareness among this socially disadvantaged community in Bangladesh.

Conclusion

Despite antibiotics being a prescription-only medicine in Bangladesh, this study identified the contributing factors of self-medication. Furthermore, various contributing factors surfaced, such as knowledge, practices, and misconceptions about antibiotic usage among the indigenous communities of Bangladesh. These indicate a lack of awareness and regulatory control across various sectors. Thus, raising public awareness and enforcing regulations in Bangladesh will encourage physicians, pharmacists, and patients to effectively address SMA to prevent AMR-mediated clinical issues among the indigenous population in Bangladesh.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article. Any additional data of this study can be obtained from the corresponding author (AM) upon request.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).

Ethics approval

This study involves human participants and was approved (Ref: 1728) by the Institutional review board of 250-bedded General Hospital, Chattogram, Bangladesh. Participants gave informed consent to participate in the study before taking part.