Article Text
PDF
Abstract
Objectives This study aims to assess self-reported health complaints and healthcare-seeking behaviours in a rural population of Bangladesh. These factors are crucial for understanding health challenges and designing effective healthcare services in rural areas.
Design A cross-sectional survey was conducted from May to October 2021.
Setting Four randomly selected administrative districts/regions of Bangladesh.
Participants A total of 1645 rural participants aged 18 years and older.
Outcome measures The study assessed the prevalence of self-reported health complaints and healthcare-seeking behaviours.
Results Among the participants, 66% (1084 out of 1645) reported experiencing health complaints, with 80% seeking care and 20% either not seeking care or opting for self-care. Multivariable analysis revealed that participants with formal occupations (adjusted OR (aOR) = 0.609; 95% CI 0.396 to 0.938; p=0.025), those from the second (aOR=1.742; 95% CI 1.014 to 2.991; p=0.044) and fifth quintiles (aOR=1.210; 95% CI 0.726 to 2.019; p=0.465), with non-communicable disease (NCD) related complaints (aOR=5.299; 95% CI 3.673 to 7.643; p <0.001), and those living more than 5 km from healthcare facilities (aOR=1.725; 95% CI 1.040 to 2.861; p=0.034), were more likely to seek healthcare. Additionally, participants in the wealthiest quintile (aOR=1.963; 95% CI 1.080 to 3.569; p=0.027), those with non-NCD complaints (aOR=5.299; 95% CI 3.673 to 7.643; p<0.001) and those living further than 5 km (aOR=4.615; 95% CI 3.121 to 6.824; p<0.001), were more likely to seek care from skilled providers or healthcare facilities.
Conclusion A high prevalence of self-reported health complaints, particularly related to NCDs, was observed. Despite this, many participants did not seek healthcare, indicating the need to address barriers to healthcare access and improve health-seeking behaviours in rural Bangladesh.
- Health Services
- Health Services Accessibility
- Health Surveys
Data availability statement
Data are available on reasonable request. The data used and analysed during this research are not publicly available due to ethical restrictions and data confidentiality. Data are available on reasonable request from researchers who meet the criteria for access to confidential data. Interested parties may contact the first author (md.kabir@monash.edu) for further inquiries in this regard.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
The survey was conducted in randomly selected four administrative districts, which cover the entire country to estimate the prevalence of self-reported health complaints and health-seeking behaviour among adult people in rural Bangladesh.
A multistage stratified cluster sampling method was used to maximise the sample representativeness of the adult population in rural settings.
Data collected were self-reported and there may be under-reporting or over-reporting of health complaints and health-seeking behaviour.
As a cross-sectional study, it cannot make causal inferences from the data.
Introduction
Bangladesh has made substantial progress in population health in recent decades.1 Reduced maternal and child mortality, total fertility, effective population control and family planning, increased life expectancy and wider immunisation coverage are considerable achievements in the health sector.2 3 However, the country is undergoing a demographic and epidemiological transition posing new challenges. The rise of non-communicable diseases (NCDs) is likely to place substantial pressures on the health system.4 NCD-related deaths and disabilities have gradually increased in Bangladesh and currently account for 71% of deaths, an increase from 45% in 2000.5 In the coming years, unless appropriate strategies and actions are in place, the rise of NCDs and infectious and neglected diseases is projected to pose substantial challenges to the healthcare system in sustainably delivering affordable, accessible and reliable healthcare services for all segments of the population.6 7
In response to the current disease burden and future predictions, a number of strategies and programmes have been introduced.8 As a common approach, healthcare services have considerably mobilised at the primary healthcare (PHC) level. The increase in healthcare services at the PHC is likely to improve the quality of healthcare services and make them more sustainable, affordable, accessible and equitable.9 It is worth mentioning that Bangladesh has a pluralistic healthcare system where various practitioners provide healthcare services and apply treatment methods and approaches.8 This pluralistic healthcare system allows an individual to seek healthcare services according to their needs and preferences.9 However, studies reported that disparities exist in accessing healthcare services due to household wealth, out-of-pocket expenses and the sociodemographical characteristics of individuals.10 11 These disparities, combined with the varying severity of health conditions, significantly affect health-seeking behaviour.12
Against this backdrop, a better understanding of morbidities, individuals' responses to and their choice of healthcare services and health system factors is important when reflecting on the rationale and effectiveness of health-related policies, strategies and actions. With regards to this viewpoint, self-reported health status, health-seeking behaviour and the use of healthcare services can be potential determinants of the population’s health and are important factors for improving public health.13 Self-reported health integrates a range of individual (eg, biological and mental) and system-level (eg, sociocontextual) factors under which individual health is determined.14 Therefore, self-reported health may offer useful information about morbidity conditions, care-seeking patterns, the disease burden, demand for healthcare services and the need for better designed healthcare services. A self-reported health status can be effective in resource-limited settings because it enables stakeholders to extract health-related information with minimal effort and resources. A great deal of health information can be included in a self-assessment questionnaire. There is a common assumption that a self-reported assessment is less reliable and focused on contextual aspects. However, self-reported health is growing in importance when it comes to public health.15
Despite the growing importance of profiling self-reported morbidity and subsequent responses to seeking healthcare, research related to the use of healthcare services in Bangladesh remains notably low.16 A better understanding of self-reported health complaints, health-seeking behaviour and the use of healthcare services is crucial for improved responses to healthcare demands and gaps in the existing service delivery mechanism. The complete profile of self-reported health complaints, healthcare-seeking patterns, health services utilisation and associated factors in a given community is crucial to better planning and managing disease burden.16 17 This study aims to fill this gap by providing a detailed profile of self-reported health complaints, health-seeking behaviour and healthcare utilisation in a rural Bangladeshi population. The findings will contribute to a better understanding of the healthcare demands and gaps in service delivery and can inform public health strategies aimed at improving health outcomes in Bangladesh and similar settings elsewhere.
Materials and methods
Study time and settings
This cross-sectional study was conducted in four administrative districts in Bangladesh—Cumilla, Jhenaidah, Rajshahi and Sylhet—from May to October 2021. The pluralistic healthcare system was observed in the studied districts, with multiple actors and healthcare providers, including the government (public sector), private operators (for-profit), non-governmental organisations, charities (not-for-profit) and donor agencies (developing partners and aids) playing roles in applying a mixed system of medical practices.8 Apart from these major formal providers, there is an extensive presence of informal healthcare providers across the districts (ie, traditional healers, faith healers, herbalists, quacks and homoeopaths). Although the organisation and delivery of the healthcare system across these districts are considerably uniform, distinctive sociodemographic characteristics, geographic features, livelihood patterns and sociocultural practices were noted.8 18–20
Sample size
Sample size was calculated by using following formula:
Here,
n=the desired sample size.
p=the proportion of the target population. We took the nationally representative data reported the age-adjusted prevalence of diabetes as 9.7%21 into account, which was the highest in Bangladesh.
p=1 p.
d=degree of accuracy desired, which is set at (0.02) 2%.
=the standard normal deviate usually set at 1.96 corresponds to the 95% CI.
The minimum required sample size was calculated to be 840. Given the nationwide scope and sociodemographic diversity of the population, the sample size was adjusted by multiplying it by a design effect of 1.522 to account for the sampling variance introduced by the multistage study design.23 This adjustment resulted in a sample size of 840*1.5=1260. Additionally, a 10% non-response rate was anticipated,23 which increased the sample size by 63, resulting in a final sample size of 1386.
Participants and sampling strategy
A multistage stratified cluster sampling was used in this study (figure 1). According to the administrative structure, Bangladesh is divided into eight administrative divisions.24 Each of the divisions is divided into several districts, and each district is divided into several subdistricts. The Bangladesh Bureau of Statistics divided the entire country into 296 718 enumeration areas (EAs) based on the latest population and housing census enumeration map.25 On average, each EA has 120 households.25 This EA list was used as a sampling frame in this study. Stepwise procedures were followed in selecting the respondents, as detailed in our protocol study.24 In the first stage, four administrative divisions were randomly chosen. Of them, four districts were randomly selected—one from each selected division, including Cumilla, Jhenaidah, Rajshahi and Sylhet. The population’s proportion of the latest census (64% in rural and 36% in urban) was considered to determine the number of participants. In the protocol, the sample size was calculated as 1386; however, we decided to increase it to 1743, considering the availability of resources and time and to increase the study power.24 26 The required smallest number of participants in any single EA was 63 in Jhenaidah. This figure was considered the maximum sampling intensity in an available EA. A total of 26 EAs were randomly chosen. In the next stage, following a systematic random sampling procedure, 63 households were selected from each EA. Applying the inclusion criteria (aged≥18 years, not pregnant and with no surgery history for the last 3 months), a single adult was interviewed in the selected household following the Kish grid method.27
Flowchart of the study participants.
Data collection procedure
Eight field enumerators administered a structured questionnaire to the sampled households. The enumerators received 1 week of training that covered the research topic, data collection instruments, administered electronic questionnaire and RedCap software.28 The purpose of the interviews was to gather essential data on the participants' sociodemographic characteristics, health complaints and healthcare-seeking behaviours. Specifically, the structured interviews were designed to collect self-reported information on health status, health-seeking actions taken in the past 30 days and the types of healthcare services used. The questionnaire used in this study was developed by the research team based on a comprehensive review of the existing literature and the contextual expertise of the researchers in the field of self-reported health complaints and healthcare-seeking behaviours. The development process involved identifying key themes from relevant studies, which were then adapted to the local context to ensure that the questions were culturally appropriate and relevant to the population under study. A plain English questionnaire was developed and then translated into Bengali (the local language). The Bengali version was retranslated into English to check the consistency state between versions. A pilot test of the questionnaire was conducted to assess its clarity, reliability and effectiveness in capturing the intended data in the study context. The necessary changes were made in the final version based on the feedback to ensure that the final version of the questionnaire was well suited to the study objectives and target population. The interview was conducted in Bengali.
Outcome measures
The prevalence of self-reported health complaints and healthcare seeking was the outcome of interest. Healthcare seeking was defined as the state of action taken by the individuals who reported health complaints, represented as a binary variable indicating whether they sought care (of any type) or opted for self-care. The survey included four possible response options for questions assessing the outcomes of interest: ‘yes’, ‘no’, ‘do not know’ and ‘not applicable’. For the purposes of analysis, we focused on the binary ‘yes’ and ‘no’, responses, as these directly correspond to whether or not the participant sought healthcare. The ‘yes’ and ‘no’ responses were coded as binary variables to simplify the analysis and to directly assess the key outcome of interest: healthcare-seeking behaviour. Among those who sought care, we further classified the source of care into skilled healthcare providers/facilities versus others (unskilled providers, self-care, etc). The survey collected self-reported healthcare-seeking data for the past 30 days to minimise the chance of recall bias, as reported in previous studies in Bangladesh and similar settings elsewhere.16 29 Previous studies reported that 30 days of self-reported information would likely provide more accurate information than a longer time period would. Healthcare-seeking behaviours were categorised as ‘sought care’ and ‘did not seek care’. When an individual reported taking no actions or that they were engaged in a treatment of their own choice without following any recommendations from healthcare providers to remedy their illness, this was defined as ‘did not seek care’. In the Bangladesh context, self-management has been a common practice for a long period of time, with individuals using various home remedies (creed and herbal) to treat their illnesses.8 30 The no-care and self-care responses were merged into one category, considering the low response rate. Beforehand, a likelihood ratio test was performed to test the feasibility of combining these two variables. When a person sought care from a semiqualified (eg, drug shopkeeper, village doctor, traditional and faith healer or homoeopath) or qualified healthcare professional (eg, general physician, nurse or specialist). The explanatory variables were chosen based on the available literature, relevant healthcare-seeking models, researchers’ expertise in the relevant fields of study and the contextual factors (eg, availability of healthcare providers and healthcare facilities in the study areas). Healthcare-seeking behaviours were associated with age, sex, education, religion and the socioeconomic status of the respondents.31 Aligning with the previous studies and contextual factors, we included a range of explanatory variables: age (in years and categorised into three groups: <40, 40–60 and≥60), sex (male or female), education (no formal education, primary school (I–V grade) or secondary school or above), religion (Muslim or other), marital status (married or unmarried), occupation (informal or formal), type of illness (NCD-related complaints or non-NCD-related complaints), location (as administrative districts), place of residency (rural or urban), distance from the facilities (≤5 km or >5 km) and socioeconomic status (as asset quintiles). Asset quintiles were obtained by combining household belongings following a principal component analysis.16
The following household belongings were factored in: land (yes or no); electricity or solar panel (yes or no); water source (yes or no); sanitary toilets (yes or no); television, radio or mobile phone (yes or no); refrigerator (yes or no); computer (yes or no); furniture, such as chairs, tables, and a bedframe (yes or no); motorbike or easy bike (yes or no); van or rickshaw (yes or no) and cooking fuel (wood, crop residue, dung cake, coal, charcoal, kerosene, electricity, liquid gas or biogas).
Statistical analysis
Data analyses were conducted using SPSS (version 24). The data were checked to fix the errors and missing values before commencing analysis with the SPSS software. The prevalence of self-reported health complaints, healthcare-seeking behaviour and the use of health services was reported as percentage with 95% CIs. A multivariable logistic regression was performed to assess factors associated with healthcare-seeking behaviours and types of healthcare providers or facilities. The logistic regression analyses reported the adjusted ORs (aORs) with 95% CIs. The relationships between the predictor variables were assessed, and the two-way term interactions, which were found significant at p<0.05, were included in the multivariable model. Similar to a backward elimination method, the predictors that were found non-significant at p<0.05 were dropped individually, and the resultant models were compared using a goodness-of-fit test until further improvements could be established. A similar process was followed to develop the final multivariable model for all outcome variables. Collinearity was checked and removed from the model if the r value was 0.70. If the p value is 0.20 or more for an independent variable in a crude or unadjusted model, it will be removed from the final multivariable model. The final models were reported with aORs, 95% CIs and p values.
Patient and public involvement
Patients were not involved in this study.
Results
Characteristics of the participants and self-reported health complaints
Table 1 shows the characteristics of the sampled participants. Data were collected from 1645 participants, with a response rate of 94%. The non-response rate was either due to the refusal to respond or the absence of household members. Among the participants with complaints, 871 (80%) sought care and 213 (20%) sought no-care or performed self-care. This sample subset has been included in the analysis as they engaged in healthcare seeking. Among the participants, 41.4% were aged between 40 and 60, 52.9% were female, 34.2% had a secondary education and above, 91.2% were married, 84.4% were in non-formal occupations and 92.0% were Muslim. Nearly, two-thirds (36.7%) of the participants had NCD-related complaints, and 63.3% had non-NCD-related complaints. NCD-related complaints included diabetes mellitus, hypertension, heart problems, stroke, cancer, kidney disease and chronic obstructive pulmonary disease/respiratory problems, while any illness/health problems other than those mentioned above were regarded as non-NCDs. Around 68.1% had access to healthcare facilities within 5 km. The bivariate analysis showed a positive association between ‘sought no-care or self-care’ with younger people (aged<40 years) (p=0.005), male (p=0.004), being married (p<0.001), having NCD-related complaints (p<0.001), availability of healthcare facilities within 5 km (p=0.013) and residence in Cumilla (p<0.001).
Characteristics of the sample participants by healthcare-seeking status
Table 2 shows the result of the regression analysis and the model specification. This analysis examines whether a respondent sought any form of healthcare, regardless of the type of healthcare provider (ie, skilled or unskilled) or chose self-care. Respondents with formal occupations (aOR=0.609; 95% CI 0.396 to 0.938; p=0.025), from the second (aOR=1.742; 95% CI 1.014 to 2.991; p=0.044) and richest quintiles (aOR=1.210; 95% CI 0.726 to 2.019; p=0.465), with non-NCD-related complaints (aOR=5.299; 95% CI 3.673 to 7.643; p <0.001), and who lived a distance of >5 km away from healthcare facilities (aOR=1.725; 95% CI 1.040 to 2.861; p=0.034), were more likely to seek healthcare.
Predictors of healthcare-seeking status (n=1084)
Table 3 shows the predictors for using skilled care providers/healthcare facilities. Respondents with fifth asset quintiles (aOR=1.963; 95% CI 1.080 to 3.569; p=0.027), non-NCDs-related complaints (aOR=5.299; 95% CI 3.673 to 7.643; p < 0.001) and a distance>5 km from the healthcare facility (aOR=4.615; 95% CI 3.121 to 6.824; p < 0.001) were more likely to seek healthcare from skilled care providers/healthcare facilities.
Predictors of the type of facilities/providers (n=871)
Discussion
This study investigated self-reported health complaints, healthcare-seeking behaviour and use of health services in rural Bangladesh. We found that a large proportion of participants (1084 out of 1645, 66%) reported having some degree of illness. A previous study in Bangladesh resulted in around 75% self-reported illnesses among the older population (aged 65 years or above), which is higher than our study.32 However, studies in neighbouring countries of Bangladesh have reported lower prevalence rates of morbidities. For instance, Poudel et al found that 48.3% of the elderly in Nepal were affected by pre-existing chronic conditions.33 In contrast, studies in India indicated that the prevalence of at least one morbidity among individuals aged 60 years and older ranges from 84% to 88%.34 35 This is likely due to people of advanced age having an increased risk of morbidities.33 As such, studies showed that the multimorbidity increased with age.36 37 Among the reported complaints, over one-thirds of the participants had NCD-related complaints. Previous studies showed an increasing trend of NCDs and associated risk factors in recent years in Bangladesh, with varying prevalence based on age, income, sex, ethnicity and geographical location.38 39 As such, studies have reported that hypertension varied between 15.9% and 30%, diabetes between 5% and 34.9%, cardiovascular diseases between 1% and 21% and chronic kidney diseases between 12.8% and 26.0%.38 40–44 Our study showed that more than one-thirds of illness complaints were NCD related, which is consistent with findings from similar studies in comparable contexts.45–47 Although we did not report each of the major NCDs separately due to the scope and aims, we assume that the combined prevalence of NCDs is likely to be consistent with these studies.
Our study showed that a participant with a non-NCD-related complaint sought care compared with the NCD-related complaints. Due to the nature of the study, we cannot adequately explain the reasons for this varying pattern of healthcare seeking. However, a few studies have reported that healthcare seeking is influenced by a range of factors, including supply-side factors (eg, health facility, care providers and cost), disease severity and psychosocial and individual characteristics.45–50 NCD-related illnesses often require continued medication support, well-equipped healthcare facilities and trained healthcare providers. These factors may not be readily available in rural settings and can involve high treatment costs, leading to avoidance of healthcare.17 45 46 Illness severity also influences healthcare-seeking behaviour, as people may not be prompt to seek care unless the illness disrupts their daily lifestyle.12 NCD-related illnesses may cause little or no impact on the daily activities of the person, which may lead to delayed or avoidance of healthcare seeking.51
Unlike the previous studies mentioned, our study found that educational entertainment did not have a significant influence on healthcare seeking in Bangladesh and in similar contexts.52 53 We suggest that a further qualitative study investigate the extent that educational entertainment influences health decision-making. Socioeconomic position (wealth quintile) was not found to be significantly associated with illness complaints or seeking of healthcare. It is generally assumed that people from affluent segments of society are more likely to develop illnesses and seek healthcare. Our findings showed that more than one-thirds of illnesses were reported to be NCD related. Possibly due to the epidemiological and demographic transitions, a higher proportion of people from relatively lower socioeconomic positions are developing illnesses. In particular, epidemiological and demographic transitions in NCD progressed; the change of developing NCDs among low-income groups rapidly increased.54 However, the findings show that the wealthiest group with NCD-related illness complaints who lived a greater distance from the healthcare facilities had higher odds of seeking skilled providers or facilities. This may be due to the fact that people belonging to the richest quintile had more reliance on private healthcare facilities, which are located in the urban centre or local township. People from the low-income group often rely on unskilled care providers, including the highly prevalent drug outlets (locally known as pharmacies), informal care providers and the practice of self-medication.55 NCD-related services and trained providers are less likely to be available in rural settings, so people with NCD-related health complaints have a greater likelihood of seeking care from privately operated facilities, usually located in townships or district headquarters or far away, to ensure quality care.
Limitations of the study
This study has certain limitations. This was a cross-sectional study, so it was unable to establish a temporal relationship between the prevalence of self-reported health complaints and the healthcare-seeking behaviour of the participants. A possible limitation is that self-reported data may be subjected to recall bias. The findings of this study were based on data from rural settings; therefore, they may not be easily representative of other populations. Notwithstanding these limitations, the findings offer important insights into the state of self-reported health complaints, healthcare-seeking behaviour and the use of healthcare services, which have certain implications for designing better health planning.
Conclusion
This study sheds light on the self-reported health complaints, health-seeking behaviours and healthcare utilisation patterns among people in rural Bangladesh. A large proportion of participants reported health complaints, with over one-thirds of these complaints being related to NCDs, highlighting the growing burden of NCDs in rural settings. Despite this, 20% of participants did not seek care, even though healthcare services were available within a reasonable distance for most individuals. The factors associated with seeking healthcare included the presence of non-NCD-related complaints, formal employment, higher socioeconomic status and access to healthcare facilities. These findings suggest that barriers to healthcare seeking, such as affordability, awareness and quality of care, continue to persist, despite efforts to improve PHC services. It is clear that tailored interventions are needed to improve healthcare access, particularly for NCD management, and to encourage more proactive health-seeking behaviours. Moreover, addressing the accessibility of skilled care providers and improving public awareness about NCDs are crucial for optimising healthcare utilisation and achieving better health outcomes in rural areas.
Supplemental material
Data availability statement
Data are available on reasonable request. The data used and analysed during this research are not publicly available due to ethical restrictions and data confidentiality. Data are available on reasonable request from researchers who meet the criteria for access to confidential data. Interested parties may contact the first author (md.kabir@monash.edu) for further inquiries in this regard.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants. The project has been approved by Monash University Human Research Ethics Committee (Project ID: 27112) and the Bangladesh Medical Research Council (BMRC) (Ref: BMRC/NREC/2019-2022/270). This study was performed in line with the principles of the Declaration of Helsinki. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
The authors would like to express their gratitude to the participants of this study. The authors are also thankful to the data enumerators.
References
Footnotes
X @Karim
Contributors AK is the review guarantor. AK, MNK and BB conceived and designed the study. AK developed the data collection tools, the data collection activities and coordinated the field operations. AK prepared the first draft of the manuscript. NK and BB revised the manuscript. BB provided overall stewardship. The final manuscript has been read and approved by all the authors.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.