Prevalence of Head Injury in India: A Systematic Review and Meta-analysis of Cross-sectional Community-based Surveys

Abstract

Background

Head injury (HI) causes a wide range of psychological symptoms, such as cognitive deficits and mood changes, leading to significant disability. Although a few studies have explored the prevalence of HI in India. However, none of the meta-analytical studies reported the estimation of the prevalence of HI in India based on cross-sectional community surveys.

Summary

A systematic review search was conducted on journal hosting sites, namely EBSCOhost, Google Scholar, Ovid, PubMed, and Web of Science, complying with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Cross-sectional Indian studies reporting the prevalence of HI in India were included in the present study. The risk of bias was assessed using the Risk of Bias Tool for Prevalence Studies. The meta-analysis was conducted using ‘meta’ and ‘dmetar’ packages through R software (4.4.0). A total of 17 studies surveyed 85,720 people, of whom 921 suffered from HI. The results showed an HI prevalence of 25.7 per thousand people in India. Subgroup analysis revealed a higher prevalence of HI among females (91.1 per 1,000) in comparison to males (85.1 per 1,000). Additionally, subgroup analysis of domicile revealed a higher HI prevalence among those from urban areas (29.5 per 1,000) than among those from rural areas (22.4 per 1,000).

Key Message

This study highlights the substantial burden of HI in India and the associated disparities in gender and domicile. The findings will be helpful to health professionals and policymakers in making informed decisions regarding the preventive aspects of HI in India.

Introduction

Head injury (HI) is a mechanical insult to the scalp, skull, or brain that may result from contact with a physical force or from sudden acceleration or deceleration movements. ¹ HI may be classified into three types, namely mild, moderate, and severe, according to predetermined scores from the Glasgow Coma Scale (GCS). It may lead to potential impairments in a variety of symptoms, such as physical, emotional, and cognitive sequelae. ² Individuals with HI may develop neurodegenerative diseases, leading to permanent disability in later life.^3–5 Research shows a high global incidence of HI in regions like Southeast Asia and the Western Pacific. ⁶ HI is found to be the most prevalent in low- and middle-income countries (LMIC) in comparison to high-income countries. ⁷ Despite this, a marked scarcity of high-quality community-based studies exists in LMIC due to issues such as underreporting of cases, absence of clear case definitions and inadequate reporting mechanisms.^{8, 9}

In the recent decade, India has experienced rapid urbanisation and motorisation, leading to a rise in HI cases from road traffic accidents (RTA).^{10, 11} Additionally, this situation is further exacerbated by factors such as poor compliance with road safety regulations (e.g., helmet use and adherence to traffic signals). ¹² Furthermore, a delay in the first line of treatment and rehabilitative efforts contributes to poorer outcomes for HI survivors. A global burden of disease study ¹³ found that over a quarter of the world’s trauma deaths occur in India, of which 80% were caused by RTA. Therefore, consideration should be given to this large percentage of people.

A review of the literature revealed only a few studies focusing on India in their investigation of the prevalence of HI. For example, Damara ¹⁴ conducted a meta-analysis on mortality due to HI in LMICs, including only two Indian studies. Similarly, Khan ¹⁵ studied the prevalence of HI in older adults in LMICs, but the focus was limited to Tamil-speaking populations. None of the above-mentioned studies employed community-based surveys to study the prevalence of HI in India. For instance, a study by Damara et al. ¹⁴ was on hospital-based surveys only. Additionally, there is a dearth of meta-analytical studies conducted to estimate the prevalence of HI in India, especially using community-based cross-sectional surveys. Therefore, the aim of this study was to perform a systematic literature review and perform a meta-analysis to estimate the pooled prevalence of HI in India based on community-based cross-sectional surveys, and further to examine the different socio-demographic characteristics associated with the occurrence of HI.

Methods

This study adhered to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. ¹⁶

Information Sources and Search Strategy

A literature search was conducted in the Ebscohost, Google Scholar, Ovid, PubMed, and Web of Science databases, published up to February 2024. Studies in the English language were eligible for screening. The search strategy included the following keywords:

‘Epidemiology’ OR ‘Prevalence’ OR ‘Incidence’ OR ‘Risk’ OR ‘Burden’ OR ‘Case fatality’ OR ‘Mortality’, AND ‘Concussion’ OR ‘Brain injur*’ OR ‘Head injur*’ ‘Traumatic brain injur*’ OR ‘Craniocerebral injur*’ OR ‘Contusion’ OR ‘Laceration’ OR ‘Skull fracture’ OR ‘Axonal injur*’ AND ‘India’.

Eligibility Criteria

To be eligible for inclusion, studies must include (a) patients with injury to the head (b) studies carried out in India (c) studies reporting HI due to RTA, workplace injury and fall, (d) community-based cross-sectional survey conducted in the general population and (e) studies providing numerical data on the total sample surveyed and the identified cases of HI. Articles were excluded if they reported HI due to non-traumatic causes, such as poisoning and stroke. Articles reporting combined injuries to the mandible or spine in addition to the head were excluded. Finally, studies were excluded if they reported HI in a particular population or due to an event, such as prisoners, victims of abuse, and disaster victims.

Data Selection and Extraction Process

The titles and abstracts of the articles were screened independently by two authors (SA and JR) according to the eligibility criteria. Duplicate articles were identified manually, and full-text screening was followed by the exclusion of the final article. The name of the author(s), year, location, study design, study period, measurement tool, domicile (urban/rural), injury site, total sample size, and number of HI cases were extracted from the article and entered into an Excel sheet. Furthermore, the proportion of individuals with HI relative to the total surveyed population was extracted for the meta-analysis.

Risk of Bias

The Risk of Bias Tool for Prevalence Studies by Hoy and colleagues was used to assess the risk of bias. ¹⁷ This tool consists of 10 items for evaluating the external and internal validity of the included studies. The interpretation of the tool’s scores followed Macaulay’s recommendations, with scores below five categorised as high risk of bias, scores from 6 to 7 as moderate risk of bias, and scores from 8 to 10 as low risk of bias. ¹⁸

Data Synthesis

The meta-analysis was performed using R software (4.4.0) employing ‘dmetar’, ¹⁹ ‘meta’, ²⁰ and ‘xlsx’ ²¹ packages. The random-effects model was chosen for conducting the current meta-analysis due to heterogeneity across the reported findings on the prevalence of HI in India. Furthermore, the ‘metaprop()’ function was used to estimate the prevalence of HI. Furthermore, a heterogeneity analysis was performed for each demographic comparison. Forest plots of the included studies were generated using the ‘forest’ function. The visual representation for publication bias was curated using Egger’s test through the ‘regtest()’ function. ²²

Results

Study Selection

A total of 1,758 articles were identified from databases. Ebscohost, Google Scholar, Ovid, PubMed, and Web of Science (see Figure 1 for the PRISMA flow diagram). A total of 233 duplicate records were screened. Further, articles were screened based on the title and abstract search. Further, a search was refined on the basis of full-text search based on pre-set inclusion and exclusion criteria. Ninety-one papers were excluded, of which 40 articles did not provide data on HI, 21 articles focused on veterans of war belonging to the United States and United Kingdom, 12 studies were review studies or systematic reviews, 12 studies belonged to other South Asian countries, and 6 studies were not community-based surveys.

Figure 1. PRISMA Flowchart Demonstrating the Study Selection Process.

Risk of Bias Assessment

Table 1 depicts the interpretation of the risk of bias of each included study. The results indicated that 35% of the included studies were scored as low risk of bias, 65% of the included studies were scored as moderate risk of bias, and none of the included studies were scored as high risk of bias.

Table 1. Source Studies in Meta-analysis of TBI in India.

Author Name	Year	Location	Study Design	Study Period	Case Ascertainment	Site	Total	Cases	Risk of Bias
Babu et al.	2022	Bangalore, Karnataka	Cross-sectional study	*	Custom-made questionnaire	Head and Neck	675	24	Low risk
Banerjee et al.	2016	Hooghly district, West Bengal	Cross-sectional study	11.2015–02.2016	Interview, Parent Supervision Attributes Profile Questionnaire (PSAPQ)	Head	163	13	Moderate risk
Basak et al.	2020	Siliguri, West Bengal	Cross-sectional study	05.2016–04.2017	Operational Definition, Custom-made questionnaire	Head and Neck	780	70	Moderate risk
Dandona et al.	2010	Godavari district, Andhra Pradesh	Cross-sectional study	10.2007–05.2008	Interview, ICD-10	Head	3686	12	Moderate risk
Dandona et al.	2006	Hyderabad, Telangana	Cross-sectional study	04.2004–11.2004	Interview, Custom-made questionnaire	Head and Neck	34,696	128	Moderate risk
Das et al.	2020	Purba Bardhaman district, West Bengal	Cross-sectional study	07.2018–06.2019	Operational Definition, Interview	Head	555	2	Low risk
Gosalia et al.	2019	Rajkot, Gujarat	Cross-sectional study	01.2014–12.2014	Operational Definition, Interview, Custom-made questionnaire	Head	2367	39	Low risk
Inbaraj et al.	2017	Vellore, Tamil Nadu	Cross-sectional study	03.2013–08.2013	Interview, Custom-made questionnaire	Head	1600	16	Low risk
Indira et al.	2016	Chidambaram, Tamil Nadu	Cross-sectional study	11.2015–08.2016	Operational Definition, Interview, Custom-made questionnaire	Head	200	0	Moderate risk
Khan et al.	2015	α	Cross-sectional study	2003–2007	Operational Definition, Interview	Head	2003	150	Moderate risk
Kumar et al.	2014	Delhi	Cross-sectional study	03.2013–04.2013	Operational Definition, Custom-made questionnaire	Head	192	2	Moderate risk
Lahare et al.	2023	Chhattisgarh	Cross-sectional study	01.2022–06.2022	Interview, Custom-made questionnaire	Head	312	32	Moderate risk
Pathak et al.	2020	Ujjain, Madhya Pradesh	Cross-sectional study	*	Operational Definition, Interview, Training Educating and Advancing Collaboration in Health on Violence and Injury Prevention (TEACH-VIP 2)	Head	6308	300	Low risk
Rajak et al.	2022	Burnpur, West Bengal	Cross-sectional study	11.2019–03.2020	Operational Definition, Custom-made questionnaire	Head	505	37	Moderate risk
Tripathy et al.	2015	Chandigarh	Cross-sectional study	03.2011–10.2012	Interview, Hindi Mini Mental Status Examination (HMSE), Physical examination	Head and Neck	300	9	Low risk
Verma & Tewari	2004	Delhi	Cross-sectional study	*	Interview, Operational Definition	Head	30,554	76	Moderate risk
Zaidi et al.	2013	Aligarh	Cross-sectional study	08.2006–07.2007	Custom-made questionnaire	Head	824	11	Moderate Risk

Notes: All included studies are mentioned in alphabetical order.

* denotes study period not mentioned; α denotes location not mentioned.

Prevalence of HI in India

The present meta-analysis included 17 studies from India, encompassing a total surveyed population of 85,720, of which 921 sustained HI.^{15, 23–38} The estimated prevalence of HI in India was found as 25.7 per 1,000 population (95% CI: 12.5 to 43.2 per 1,000 population). The heterogeneity analysis (Q = 1391.93, df = 16, p < .001) and I ² value (98.9%) suggest the presence of significant heterogeneity across the studies, reflecting varying patterns of HI prevalence across India (see Figure 2 for more details).

Figure 2. Forest Plot Representing the Studies Included in the Prevalence of HI in India.

Prevalence of HI as per Socio-demographic Characteristics

Table 2 presents the prevalence of HI in India according to sociodemographic characteristics. A comparison of 12 urban and 6 rural-based studies (Figure 3) revealed that those residing in urban areas suffered more from HI (29.5 per 1,000 population, 95% CI: 12 to 54.1 per 1,000 population) than their rural counterparts (22.4 per 1,000 population, 95% CI: 5 to 50.7 per 1,000 population). Subgroup analysis indicated no significant difference between rural and urban populations (Q = 0.18, df = 1, p = .67).

Table 2. Prevalence of TBI as per Domicile, Age and Sex.

	No of Studies	Prevalence	95% CI	Tau 2	Q	I 2	Subgroup Differences
Domicile
Urban	12	0.0295	0.0120, 0.0541	0.0112	734.05	98.5%	Q = 0.18, df = 1, p = .67
Rural	6	0.0224	0.0050, 0.0507	0.0083	129.15	96.1%
Sex
Male	4	0.0851	0.0032, 0.2525	0.0535	199.14	98.5%	Q = 0.01, df = 1, p = .94
Female	4	0.911	0.0074, 0.2451	0.0436	104.94	97.1%

Figure 3. Forest Plot Representing the Included Studies as per Domicile.

Furthermore, upon comparing the prevalence of HI between males and females (Figure 4), the prevalence among females (91.1 per 1,000 population, 95% CI: 7.4 to 245.1 per 1,000 population) was found to be slightly higher than that among males (85.1 per 1,000 population, 95% CI: 3.2 to 252.5 per 1,000 population). Subgroup analysis showed no significant difference between the genders (Q = 0.01, df = 1, p = .94).

Figure 4. Forest Plot Representing the Included Studies as Per Gender.

Publication Bias

The funnel plot is a visual representation of the potential existence of publication biases related to the inclusion of studies for the current prevalence estimation. Studies with small sample sizes and non-significant results have very little chance of being published, so it is also uncommon for those studies to be excluded from the current meta-analysis. The asymmetry in the funnel plot also suggests the presence of possible publication bias in the study (Figure 5). The Egger’s regression test was used to assess the symmetry in the funnel plot, using the regtest() function of ‘dmetar’ package. ²² The regression model’s intercept, β0 = 7.59 (95% CI: 2.5–12.67), was found to be statistically significant (t = 2.93, p = .01). This result implies that there is publication bias present in this study and the standard errors of the studies that are part of the funnel plot are asymmetric.

Figure 5. Funnel Plot Representing Publication Bias Among the Included Studies.

Discussion

The present systematic review and meta-analysis is the first comprehensive research record estimated pooled prevalence of HI in India. The estimated prevalence of HI in India is about 25.7 per 1,000 persons, and women sustained HI more than men. The estimated prevalence of HI in this study is consistent with the prevalence rates reported for other LMICs. For instance, the prevalence of HI was found to be 49 per thousand population in Puerto Rico, ¹⁵ whereas in Ghana, it was reported as 33 per thousand population ³⁹ followed by the Republic of China with 4.4 per thousand population. ⁴⁰ In contrast, the HI prevalence in the United States of America ranges between 120 and 182 per thousand persons.^41–43 Similarly, a large community-based survey conducted in Ethiopia reported an HI prevalence of 397 per thousand population. ⁴⁴

The lower estimated prevalence of HI in India compared to other developed western countries can be attributed to the fact that Indian states did not have a record of a comprehensive health care system for various types of head injuries, including the milder forms of HI caused by RTAs, violence, falls and natural disasters. ⁶ Although HI in India is mainly attributed to factors such as urbanisation, densely packed infrastructure, RTAs, and interpersonal violence etc. ⁹ Surprisingly, HI in female participants were more than in their male counterparts, which is contradictory to a study conducted in India, which reported that men sustained HI more than women. ¹³ There are manifold reasons that may be attributed to these contradictory findings. For instance, while performing sensitivity analysis in the present study, it was noted that there are two influential studies comprised for overall prevalence estimation of HI and gender-based subgroup comparisons,^{15, 35} and both of these influential studies focused mainly on older (over 60 years) and younger (under 18 years) participants. In these age groups, HI is often caused by falls rather than by RTA, therefore, women may suffer more from HI than men. Further, women who participate in sports and activities are at greater risk of being injured by HI than men. ⁴⁵ Furthermore, large proportions of survivors of domestic violence and intimate partner violence are women who may sustain head injuries more than men. ⁴⁶

A comparison of rural and urban populations showed that urban residents sustained HI more than rural residents. This finding was consistent with the findings of previous research in India.^{12, 47} The urban residents may be at a higher risk of HI than their rural counterparts because of denser infrastructure, leading to a higher number of RTAs. ⁴⁷ Moreover, urban HI survivors have easy access to health care services compared to their rural counterparts, which increases their chances of being screened and identified as having a case of HI than the rural ones. ¹² The present finding is contradictory to the findings of a meta-analysis that found more HI was experienced by people in rural areas than by those in urban areas. ⁴⁸ Although most of the studies included in this meta-analysis were carried out in developed countries, ⁴⁸ where rural areas have a higher incidence of road-related brain injuries than rural India; therefore, these contradictory findings may occur in the present study.

Significant heterogeneity was observed in the included studies for this prevalence estimate as well as possible publication bias was present in this study. The heterogeneity across findings of included studies could be attributed to the use of different psychometric tools and methods for diagnosing cases of HI in the community by the different researchers.

This study is pioneering in estimating the prevalence of HI in India based on community-based cross-sectional surveys, offering both theoretical and practical contributions. For instance, findings of the current study will help policymakers to make informed decisions regarding implementing rehabilitation services. As even milder forms of HI can lead to various cognitive and affective consequences, which may persist for several months to years.^{49, 50} Therefore, there is a critical need for comprehensive treatment and care services for TBI patients in India. However, such services remain fragmented and inadequate, particularly in the management of neuropsychological symptoms associated with HI. ¹² The present study also highlighted one of the significant challenges related to the lack of well-documented health care records for non-RTA causes of TBI. This gap in documentation hampers accurate estimation of the true prevalence of head injuries across the population; therefore, findings may not fully capture the actual burden of HI in India.

Although this meta-analytical study is the first to attempt to estimate the prevalence of HI in India. However, a few limitations must be noted while interpreting the present findings, first, since the included studies were community-based, the primary method of data collection in majority of the studies were custom-made questionnaires and interviews, only a few studies used standardised measurement tools for assessing HI, such as the GCS or neuroimaging scans, therefore, this was also a major limitation of the study. Furthermore, few studies have been conducted in geographically limited areas, such as small villages or particular districts. Therefore, findings have limited generalisability to the broader population of HI patients.

Conclusion

India faces a significant public health challenge with HI, similar to the rest of the world. This study revealed that females and urban residents were at a higher risk of HI than males and rural residents in India. This study highlights the need to strengthen road safety legislation, improve trauma health and rehabilitation services for patients with HI in India.

Statement of Ethics

There were no human participants in this study; therefore, ethical approval was not required.

Abbreviations

HI: Head injury

LMIC: Low- and middle-income country

PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines

RTA: Road traffic accidents

Authors’ Contributions

Saishree was responsible for the original draft preparation, review and editing, methodology development, data curation, and conceptualization of the study. Jay Kumar Ranjan contributed to formal analysis, investigation, methodology, supervision, and manuscript review and editing. H.S. Asthana provided support in supervision, methodology, and manuscript review and editing. Nityanand Pandey contributed to supervision, methodology, and critical review and editing of the manuscript. All authors have read and approved the final manuscript.

Informed Consent

Informed consent was not applicable.