Assessment of quality of life after traumatic brain injury in adults from Armenia, Georgia, and Moldova using EQ-5D-5L

Abstract

Purpose Traumatic brain injury (TBI) occurs more frequently in low and middle-income countries (LMICs) than in high-income settings, yet data on health-related quality of life (HRQoL) from these regions remain scarce. This study assessed HRQoL outcomes in adult TBI patients in Armenia, Georgia, and Moldova. Methods Between March and September 2019, TBI data were collected using a standardized hospital-based registry in one trauma hospital per country. Demographics, injury characteristics, and symptoms were recorded; HRQoL was assessed at discharge using EQ-5D-5L. Results 386 patients were included, most with mild TBI (GCS 13–15). Falls (51%) and road traffic incidents (29.8%) were the leading causes, with males predominantly affected. HRQoL scores varied across countries, with higher values in Moldova and Armenia compared to Georgia. A significant negative correlation (r = − .201, p < .001) was observed between GCS and HRQoL, indicating that mild TBI patients did not always report higher HRQoL compared to those with more severe injuries. In multivariable regression, age, country, and concomitant injuries independently predicted lower EQ-5D index scores, while sex, GCS, and length of stay were not significant. Conclusion These findings highlight the need for post-injury HRQoL assessment and standardized hospital-based TBI registries to guide rehabilitation efforts in LMICs.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-26065-2.

Introduction

Traumatic brain injury (TBI) is a major source of disability and death worldwide, with new annual cases estimated between 27 and 69 million^1,2. The incidence of TBI is increasing, and common mechanisms of injury are road traffic incidents, falls, sports/leisure activities, and violence. When these traumatic experiences lead to a TBI diagnosis, the outcomes can be complex, as in addition to the emotional impact of the injury, the TBI itself can lead to long-lasting physical and cognitive symptoms. The physical and cognitive symptoms experienced by individuals with TBIs also pose a burden to families and caregivers. Research shows that cognitive and emotional well-being can be impacted regardless of the severity of the TBI^3–5. Still, few studies have examined the association of perceived quality of life related to TBI severity, especially in lower and middle-income countries.

Assessing the health-related quality of life (HRQoL) in patients with a TBI is crucial for understanding the injury’s long-term effects and developing effective interventions to improve the patient’s overall quality of life^6,7. HRQoL measures provide essential information about patients’ perceptions of their health status, including their physical, emotional, and social well-being⁶. Prior research shows that TBI diagnoses and symptoms vary widely, but the factors driving this variation and their impact on patients’ perceived quality of life are not yet well understood, limiting the ability to provide optimal care.

Over 80% of cases of TBI occur in low-income and middle-income countries (LMICs)³, and LMICs experience three times the TBI rate of high-income countries (HICs)¹, yet over 80% of the research on TBI is conducted in HICs^3,4. In most LMICs from Eastern Europe and Central Asia, data on TBI are scarce. Existing evidence shows that LMICs face challenges in collecting TBI data in a standardized manner and offering pre-hospital and rehabilitation care to patients due to a lack of proper equipment, protocols, and human resources^8–12. TBI is a major public health issue in the Republic of Armenia, Georgia, and the Republic of Moldova, which are classified as lower-middle to upper-middle-income countries. All three countries follow international TBI treatment guidelines and have trained medical personnel, but critical gaps exist. A fairly recent assessment found a severe shortage of pre-hospital care, specialty equipment, and rehabilitation services for brain injury patients outside major cities. There is no standard national registry for TBI, and data on long-term outcomes are essentially nonexistent. Moreover, little is known about the epidemiological characteristics of TBI and the HRQoL of patients suffering from TBI in this region, making it particularly concerning given the high incidence of TBI in Eastern Europe and the limited resources available for patients and their families^8,10. The scarce existing evidence, however, points toward similar trends in terms of the mechanism of injury, severity, and outcomes of TBI in these middle-income countries^13–16. Still, little research exists to assess the predictors of TBI-related health and HRQoL in the Republic of Armenia, Georgia, and the Republic of Moldova, this information being essential for healthcare providers, researchers, and policymakers to develop effective interventions and to evaluate the impact of these interventions on patients’ lives. Therefore, this article aims to assess TBI severity and the broad symptoms among patients with a TBI from the Republic of Armenia, Georgia, Republic of Moldova and their association with HRQoL.

Methods

Study setting and study population

This research article uses data from a larger study that examined both adult and pediatric TBI and includes adult patients between 19 and 94 from three hospitals (Emergency Medicine Institute in the Republic of Moldova, Academician O. Gudushauri National Medical Center in Georgia and Yerevan State Medical University “Heratsi Complex” in the Republic of Armenia) chosen for their volume of trauma patients. Patients aged 18 or younger were not included in this study because, based on the incidents’ circumstances, they were taken to a pediatric hospital. TBI cases were identified over the duration of six months (between March and September 2019) using an established TBI registry at local level in each of the three countries^10,13. The project and the implementation of the TBI registry were approved by the Ethics Boards of each of the partner institutions and participating hospitals before its start.

Data was collected in two phases: (1) data was extracted from medical records using the TBI registries and included all patients admitted to the partner hospitals with a TBI diagnosis during the six-month study period, and (2) patients completed the EQ-5D-5L survey at discharge to document self-reported health outcomes and HRQoL as required by the EQ-5D data collection protocol. Trained research personnel discussed with the patient and applied the EQ-5D-5L survey in their national language (Armenian, Georgian and Romanian, which is the official language of Moldova). Out of the total number of patients admitted with TBI and identified by the trained research personnel, 25.4% (N = 98) did not accept to complete the quality-of-life assessment survey (Fig. 1).

Fig. 1.

Flowchart for data collection.

Study measures

TBI registry

The study used a comparable TBI registry implemented locally in the three participating countries, which contained core TBI information adapted from existing TBI registries by contacting investigators funded by NIH and EU ¹⁰ and tailored to fit the characteristics of all three countries included in the study. The TBI registry was a novel tool developed with core variables and definitions that were then translated into the local language in each country. The process of creating the TBI registry and protocol is described in more detail elsewhere^10,13. Variables from the TBI registry explored in the present study include socio-demographic characteristics (age, sex, area where injury occurred, injury intent, means of arrival to the hospital, work-related injury, and place of occurrence of injury) and TBI-related diagnoses (Glasgow Coma Scale (GCS) score measured at 3 different times (pre-hospital GCS, GCS at the emergency department and at discharge, with the following categories: Severe GCS: 3–8, Moderate GCS: 9–12, Mild GCS: 13–15) using an open question, loss of consciousness, post-traumatic amnesia, consciousness alteration (measured using a nominal scale: yes, no, suspected, unknown), hematoma, surgery performed (measured using a nominal scale: yes, no, unknown), discharge disposition). The Glasgow Coma Scale (GCS) is positively oriented, with higher scores reflecting better neurological function, whereas the EQ-5D is likewise positively oriented, with higher values corresponding to better health-related quality of life.

Health-Related Quality of Life assessment (HRQoL) with EQ-5D-5L

HRQoL data can be analyzed and reported in multiple manners, depending on the nature of the study and its outcomes. The EuroQol Research Foundation (EuroQol) developed the tools to measure HRQoL, which is being used worldwide in different settings and for many health conditions¹⁷. The EQ-5D-5L tool was chosen for this study due to: (1) its feasibility to be administered in hospital settings; (2) it is validated and used in more than 150 languages, including the three languages used in our study; (3) the existent datasets enable comparison of outcomes across different populations; (4) previous use of the tool in other TBI studies¹⁸. Before conducting the study, the authors received permission to use the instruments, and the creators of the tools offered all the necessary materials for data collection in all three languages. EQ-5D-5L measures HRQoL across five dimensions using the EQ-5D descriptive system: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension was rated by each participant using a five-point response option ranging from “1 - no problems’’ to “5 - unable to’’. The visual analog scale (EQ-VAS) was used after the descriptive system and comprises a score from 0 (the worst health condition) to 100 (the best health condition) for patients to provide an overall assessment of their quality of life at discharge¹⁸.

Some reporting methods include value sets, EQ-5D-5L profiles, regression analysis, and EQ-VAS scores¹⁹. For this study, the EQ-5D-5L descriptive system, the score of the visual analog scale (EQ-VAS) and the index score with value sets were used. The EQ-5D index score is a measure used to assess health-related quality of life by converting responses from the EQ-5D-5L descriptive system into a simple summary value. The index value is a tool developed to explain how good or bad a patient’s health status is according to the overall preferences of a region/country. The EQ-5D index score is determined by using a formula that attaches values (weights) to possible health status, based on preferences collected from populational studies, indicating how desirable or undesirable a given health state is perceived. The correct weights start from 1, the value representing full health (i.e., 11111) and decrease based on the severity of reported problems in the 5 dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/depression). For example, if a certain health state (e.g., 12345) has an assigned value of 0.65, it means that this condition is perceived as 65% of full health¹⁷.

For this study, the Romanian value set was used, as is the common practice recommended in the EQ-5D Manual to apply a value set from a country that most closely aligns with the characteristics of the target country when a country-specific value set is unavailable, which was the case for all three countries included in the study¹⁷. Although this practice might have some limitations, when it comes to healthcare systems and access to treatment, to the best of our knowledge, this is the recommended use of the tool.

Patients who consented and were included in the TBI registry were followed up at discharge by trained hospital personnel who asked patients to complete the five-level, five-dimensional tool (EQ-5D-5L) and choose on the visual analog scale of EQ-VAS to self-assess their health-related quality of life.

Statistical analysis

Socio-demographic variables, TBI registry variables, and EQ-5D-5L tools were analyzed using RStudio, as well as computing the EQ-5D-5L descriptive system into an index score. Frequency statistics were computed for the main demographics and TBI-related variables for the entire sample. Inferential statistics were conducted using the patients who completed the EQ-5D-5L since the analysis explored the relationship between HRQoL and other factors. Correlations between the GCS score and EQ-5D-5L index were explored using Spearman’s rho. To examine independent predictors of HRQoL, we performed multivariable linear regression with EQ-5D index as the dependent variable. Covariates included age, sex, country, Glasgow Coma Scale (GCS), length of stay, and concomitant injuries. Regression coefficients (B), standardized betas (β), 95% confidence intervals, and model fit statistics were reported.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Institutional Review Board of Nicolae Testemițanu State University of Medicine and Pharmacy, Moldova (Date 15.03.2018/No. 44 to 42), Ethics Committee of Yerevan State Medical University after Mkhitar Heratsi (20.09.2018/No. N_1/2018), Institutional Review Board of National Center for Disease Control and Public Health (11.01.2018/No. 2018-002).

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Results

Socio-demographics and TBI-related symptoms of the sample

The TBI registry in all three countries included a total number of 386 adults aged between 19 and 94 years of age. Of these patients, 29 (7.5%) were from the Republic of Armenia, 156 (40.4%) from Georgia, and 201 (52.1%) from the Republic of Moldova, reflecting the relative size of the hospitals in each country and their admitted number of patients (Table 1). Out of the total 386 adults, 288 patients with TBI completed the EQ-5D-5L survey, having a 75% response rate for the HRQoL assessment (Fig. 1). Most TBIs were unintentional (86%, N = 332), followed by assault (11.9%, N = 46). The hospital in the Republic of Moldova reported the highest prevalence of assault-related TBIs (N = 32). Overall, the main mechanism of injury responsible for the TBIs was falls, followed by RTI, and most injuries were sustained at home or in a transportation area. The main mechanism of injury was falls in the Republic of Moldova (49.8%) and Georgia (57.7%), while in the Republic of Armenia, road traffic injuries were the most common (41.4%). In all 3 countries, the main type of arrival to the hospital was by ambulance (the Republic of Moldova 95.5%, the Republic of Armenia 92.8%, Georgia 59.6%) (Table 1).

Table 1.

Socio-demographic characteristics of hospital-admitted patients with TBI.

	Moldova		Armenia		Georgia		All		χ2
Characteristics	N	%	N	%	N	%	N	%
Age									31.175*
19–34	48	23.9	9	31.0	75	48.1	132	34.2
35–59	78	38.8	18	62.1	48	30.8	144	37.3
60+	75	37.3	2	6.9	33	21.2	110	28.5
Total	201	100	29	100	156	100	386	100
Mean age	51		42.44		41.79		46.67
Sex									8.904*
Male	146	72.6	17	58.6	92	59.6	256	66.3
Female	55	27.4	12	41.4	63	40.4	130	33.7
Total	201	100	29	100	156	100	386	100.0
Area where injury occurred									15.910*
Urban area	170	84.6	28	96.6	154	98.7	352	91.2
Rural area	31	15.4	0	3.4	1	0.6	32	8.3
Unknown	0	0.0	1	0.0	1	0.6	2	0.5
Total	201	100	29	100	156	100	386	100
Injury intent									9.678
Unintentional	167	83.1	25	86.2	140	89.7	332	86
Intentional self-harm	1	0.5	0	0.0	3	1.9	4	1
Assault (violence)	32	15.9	3	10.3	11	7.1	46	11.9
Other/unknown	1	0.5	1	3.4	2	1.3	4	1
Total	201	100	29	100	156	100	386	100
Work-related injury									23.409*
Yes	20	10.0	2	6.9	8	5.1	30	7.8
No	181	90.0	25	86.2	147	94.2	353	91.5
NA/Unknown	0	0.0	0	0.0	1	0.6	3	0.8
Total	201	100	29	100	156	100	386	100
Injury mechanism									28.900*
Road traffic injury	63	31.3	12	41.4	40	25.6	115	29.8
Fall	100	49.8	7	24.1	90	57.7	197	51
Assault (violence)	32	15.9	3	10.3	12	7.7	47	12.2
Struck by/or against	6	3.0	5	17.2	9	5.8	20	5.2
Other	0	0.0	2	6.9	5	3.2	7	1.8
Total	201	100	29	100	156	100	386	100
Means of arrival to the hospital									53.127*
Walk-in	9	4.5	5	17.2	41	26.3	55	14.2
Ambulance	192	95.5	34	82.8	93	59.6	309	80.1
Private/public vehicle	0	0.0	0	0.0	3	1.9	3	0.8
Police	0	0.0	0	0.0	1	0.6	1	0.3
Other	0	0.0	0	0.0	18	11.5	18	4.7
Total	201	100	29	100	156	100	386	100
Place of occurrence									74.139*
Home	58	28.9	3	10.3	51	32.7	112	29.0
Institutions	24	11.9	1	3.4	19	12.2	44	11.4
Sports and recreational areas	33	16.4	1	3.4	7	4.5	41	10.6
Streets and highways	54	26.9	16	55.2	60	38.5	130	33.7
Commercial areas	16	8.0	1	3.4	1	0.6	18	4.7
Countryside	15	7.5	0	0.0	2	1.3	17	4.4
Other/unspecified	1	0.5	7	24.1	16	10.3	24	6.2
Total	201	100	29	100	156	100	386	100

* p-value < 0.05.

Based on emergency department GCS scores, most of the injuries fell into the category of mild TBI (76.2%). The Republic of Moldova had the highest number of both moderate TBIs (24.4%) and severe injuries (11.9%) (Table 2). TBI symptoms such as headache (99.5%) and nausea (90.9%) were present in most patients. Concomitant injuries were common among hospitalized TBI patients. Thoracic injuries were most common (50.3%), followed by lower extremities (48.7%), upper extremities (47.7%) and abdominal injuries (37.3%). Significant group differences were observed across countries for all types of concomitant injuries (p <.001). Mean length of stay also differed by country, ranging from 5.22 days in Georgia, to 7 days in Armenia and 9.43 days in Moldova. Most of the patients with TBI returned home after receiving care from the emergency department (68.7%), but one-third of patients with TBI were admitted to the hospital and required more treatment. Some were transferred to rehabilitation facilities (15.7%) or admitted to acute care (10.4%) (Table 2).

Table 2.

TBI-related clinical profile of patients from the Republic of Armenia, Georgia, and Republic of Moldova.

	Moldova		Armenia		Georgia		All		χ2
Characteristics	N	%	N	%	N	%	N	%
GCS Score									24.801*
3–8 (severe)	24	11.9	2	7.4	10	6.5	36	9.4
9–12 (moderate)	49	24.4	2	7.4	4	2.6	55	14.4
13–15 (mild)	128	63.7	23	85.2	141	91.0	292	76.2
Total	201	100	29	100	155	100	383	100
Loss of consciousness									99.666*
Yes	82	40.8	11	39.3	152	97.4	245	63.6
No	80	39.8	13	46.4	0	0.0	93	24.2
Suspected	35	17.4	2	7.1	3	1.9	40	10.4
Unknown	4	2	2	7.1	1	0.6	7	1.8
Total	201	100	28	100	156	100	385	100
Post-traumatic amnesia									146.086*
Yes	54	26.9	12	41.4	156	100	222	57.5
No	101	50.2	13	44.8	0	0	114	29.5
Suspected	42	20.9	2	6.9	0	0	44	11.4
Unknown	4	2	2	6.9	0	0	6	1.6
Total	201	100	29	100	156	100	386	100
Consciousness alteration									79.378*
Yes	46	22.9	5	17.2	11	7.1	62	16.1
No	72	35.8	22	75.9	106	67.9	200	51.8
Suspected	80	39.8	1	3.4	7	4.5	88	22.8
Unknown	3	1.5	1	3.4	32	20.5	36	9.3
Total	201	100	29	100	156	100	386	100
Headache									0.854
Yes	200	99.5	25	100.0	150	99.3	375	99.5
No	1	0.5	0	0	1	0.7	2	0.5
Total	201	100	25	100	151	100	377	100
Nausea									37.737
Yes	193	96	13	56.5	135	89.4	341	90.9
No	8	5	10	43.5	16	19.6	34	9.1
Total	201	100	23	100	156	100	375	100
Vomiting									120.077*
Yes	191	95	5	27.8	51	33.8	247	66.8
No	10	5	13	72.2	100	66.2	123	33.2
Total	201	100	18	100	151	100	370	100
Balance									85.747*
Yes	128	63.7	3	15.8	151	100	282	76
No	73	36.6	16	84.2	0	0	89	24
Total	201	100	19	100	151	100	371	100
Dizziness									146.844*
Yes	59	29.4	8	40	151	100	218	58.6
No	142	70.6	12	60	0	0	154	41.4
Total	201	100	20	100	151	100	372	100
Visual									8.570*
Yes	40	52.7	0	0	7	4.6	47	12.7
No	161	47.3	17	100	144	95.4	322	87.3
Total	201	100	17	100	151	100	369	100
Fatigue									83.246*
Yes	106	52.7	8	40	151	100	265	71.2
No	95	47.3	12	60	0	0	107	28.8
Total	201	100	20	100	151	100	372	100
Light									72.598*
Yes	100	49.8	0	0	4	2.6	104	28.2
No	101	50.2	17	100	147	97.4	265	71.8
Total	201	100	17	100	151	100	369	100
Fracture of the skull									128.481*
Yes	168	83.6	12	41.4	12	7.7	192	49.7
No	33	16.4	17	58.6	144	92.3	194	50.3
Unknown	201	100	29	100	156	100	386	100
Total	165	100	194	100	387	100	746	100
Concussion									17.062*
Yes	186	92.5	22	75.9	152	97.4	360	93.3
No	15	7.5	7	24.1	4	2.6	26	6.7
Total	201	100	29	100	156	100	386	100
Concomitant injuries									234.215*
Thorax
Yes	176	87.6	5	17.2	13	8.3	194	50.3
No	25	12.4	24	82.8	143	91.7	192	49.7
Total	201	100	29	100	156	100	386	100
Abdomen									165.508*
Yes	136	67.7	0	0	8	5.1	144	37.3
No	65	32.3	29	100	148	94.9	242	62.7
Total	201	100	29	100	156	100	386	100
Upper extremities									136.089*
Yes	153	76.1	5	17.2	26	16.7	184	47.7
No	48	23.9	24	82.8	130	83.3	202	52.3
Total	201	100	29	100	156	100	386	100
Lower extremities									171.446*
Yes	162	80.6	2	6.9	24	15.4	188	48.7
No	39	19.4	27	93.1	132	84.6	198	51.3
Total	201	100	29	100	156	100	386	100
Length of stay	Mean (SD)		Mean (SD)		Mean (SD)		Mean (SD)
	9.43 (13.11)		7.04 (5.81)		5.22 (15.53)		7.55 (13.91)
Discharge Disposition									64.358*
Home	106	52.7	25	96.2	132	84.6	263	68.7
Rehab	59	29.4	0	0	1	0.6	60	15.7
Nursing home/long-term care	0	0	0	0	2	1.3	2	0.5
Acute care	35	17.4	0	0	5	3.2	40	10.4
Unknown	1	0.5	1	3.8	16	10.3	18	4.7
Total	201	100	26	100	156	100	383	100

* p-value < 0.05.

Of the 386 patients, 286 (74%) completed EQ-5D. Non-completion was significantly associated with injury severity (χ² = 81.13, p <.05). Among respondents, 84.6% had mild TBI, 13.6% moderate and only 1.7% severe. In contrast, non-respondents included 32% severe, 16.5% moderate and 51.5% mild cases (Table 3).

Table 3.

Comparison between respondents vs. non-respondents to EQ-5D on GCS severity.

	With EQ-5D		Without EQ-5D		Total		χ2
GCS Score	N	%	N	%	N	%
3–8 (severe)	5	1.7	31	32	36	9.4	81.133*
9–12 (moderate)	39	13.6	16	16.5	55	14.4
13–15 (mild)	242	84.6	50	51.5	292	76.2
Total	286	100	98	100	386	100

*p-value < 0.05.

Health-related quality of life for adult patients with TBI as assessed by EQ-5D

After establishing the socio-demographic and injury characteristics of TBI patients, we examined how these factors influence HRQoL outcomes measured using the EQ-5D-5L tool.

At the domain level (Fig. 2), the most frequently reported problems were pain/discomfort and anxiety/depression, followed by difficulties with usual activities and self-care, while mobility was least affected.

Fig. 2.

EQ-5D-5L Domains by country.

The EQ-VAS score presented substantial cross-country differences (Fig. 3). Patients in Moldova and Armenia reported high EQ-VAS scores, with medians of 90 and 93, respectively, indicating generally favorable self-assessed health. In contrast, patients in Georgia had lower scores, with a median of 50 and a wide interquartile range, suggesting both poorer perceived health and greater variability in outcomes. Outliers were observed across all groups, particularly in Georgia, underscoring the heterogeneity of patient recovery experiences.

Fig. 3.

EQ-VAS distribution by country.

The EQ-5D index values also varied significantly by country (Fig. 4). Moldova and Armenia displayed distributions clustered toward the higher end of the scale, with most patients approaching the maximum score of 1.0, reflecting fewer reported problems across EQ-5D dimensions. In contrast, Georgia demonstrated lower median index scores and a broader spread of responses, including one negative value (below 0).

Fig. 4.

Index score by country.

The detailed descriptive table is available in the Supplementary Material (Supplementary Table 1).

Index score representing quality of life

To see if there is any correlation between the GCS score registered in the Emergency Department and the quality of life reported by the patients using the EQ-5D-5L survey, we computed the index score for the scale, which indicates the quality of life measured against a population standard. The score ranges from − 0.2 to 1.0 and captures the overall quality of life, with higher scores indicating better quality of life. Only one patient (< 0.5%) reported a negative score. The distribution is right-skewed, indicating that most patients in the sample reported high levels of quality of life post-injury. On the other hand, the lower end of the distribution shows relatively few patients with low index scores. However, the presence of some negative scores highlights a subset of patients experiencing severe impairment in their quality of life. This could be attributed to factors such as the severity of the initial injury, comorbid conditions, or inadequate access to post-injury care and support. The figure representing the distribution of EQ-5D-5L index score ca. be found in the Supplementary Material (Supplementary Fig. 1).

Correlation between GCS score and EQ-5D index score

Given that the GCS is widely used to assess the neurological status in TBI patients, we explored its relationship with the EQ-5D-5L index score to determine whether injury severity correlates with self-reported health status at discharge. For this, Spearman’s rho was used to assess the relationship between the GCS scores assessed at the emergency department and the EQ-5D-5L index scores among TBI patients. The analysis revealed a statistically significant, albeit weak, negative correlation between GCS scores and EQ-5D-5L index scores (ρ = −0.29, p <.001). This suggests that higher GCS scores were associated with lower self-reported HRQoL on the EQ-5D-5L index. A sensitivity analysis applying GCS-based weights to correct for the underrepresentation of severe cases showed nearly identical association (weighted ρ = −0.28, p <.001), indicating that nonresponse bias due to underrepresentation of severe TBI cases is unlikely to substantially affect the findings (Table 4).

Table 4.

Correlation between GCS score and EQ-5D index score.

Variables	EQ-5D index score	GCS at admission
EQ-5D index score	1.000	− 0.286**
GCS at admission	- 0.286**	0.000

**p.001.N=288

The scatter plot in Fig. 5 illustrates this relationship, showing a general trend where higher GCS scores are associated with lower EQ-5D-5L index scores. The majority of GCS scores were clustered around the values of 14 and 15, with most of the patients having good neurological functions. On the same note, there were fewer cases of low index scores, suggesting a generally high quality of life. Despite the negative correlation, the spread of data points indicates considerable variability in how patients’ neurological status correlates with their perceived quality of life.

Fig. 5.

GCS Score vs. Index Score in a sample of TBI patients from low-middle-income countries*.

*6.6% of mild GCS who experienced worse quality of life scores. 1.8% of severe GCS who experienced good/positive quality of life scores.

Further analysis of outliers provides additional insight into this variability. Table 5 highlights the cases where mild TBI patients (GCS 13–15) reported poor EQ-5D-5L index scores, while a small subset of severe TBI patients (GCS 3–8) reported unexpectedly high quality of life scores. Specifically, 6.6% (N = 16) of patients with mild GCS self-assessed with worse quality of life outcomes (index score < 0.3), while 1.8% (N = 4) with severe GCS self-assessed with good quality of life outcomes (index score > 0.8). Both groups are outliers, and when using this type of analysis, the need for a more nuanced approach to assessing TBI severity is highlighted (Table 5).

Table 5.

GCS scores and index scores per category.

Index score per category	Glasgow Coma Scale score per category
	Severe (3–8)	Moderate (9–12)	Mild (13–15)
	N (%)	N (%)	N (%)
< 0.3	0 (0)	0 (0)	16 (6.6)
Between 0.4–0.7	1 (20)	2 (5.1)	45 (18.6)
> 0.8	4 (80)	37 (94.9)	181 (74.8)
Total	5 (100)	39 (100)	242 (100)

To further examine factors associated with HRQoL, we conducted a multivariable linear regression with the EQ-5D index as the dependent variable. The model, including age, sex, GCS, length of stay, country and concomitant injuries, was significant (F(6, 277) = 28.06, p <.001), explaining 37.8% of the variance in the EQ-5D index. Age (β = −0.142, p =.005), country (β = −0.606, p <.001) and the presence of concomitant injuries (β = −0.199, p =.003) were independent predictors of lower HRQoL. Sex, length of stay and GCS were not significantly associated with EQ-5D index (Table 6).

Table 6.

Multivariable linear regression analysis of factors associated with EQ-5D Index.

Predictor	B	SE	β	t	p	95% CI for B
Constant	7.369	0.585	—	12.60	< 0.001	[6.22, 8.52]
Age	−0.002	0.001	−0.142	−2.85	0.005	[− 0.003, − 0.000]
Sex	−0.017	0.020	−0.042	−0.86	0.389	[− 0.057, 0.022]
Country	−0.156	0.014	−0.731	−10.88	< 0.001	[− 0.184, − 0.127]
GCS at admission	−0.001	0.007	−0.006	−0.13	0.900	[− 0.014, 0.012]
Length of stay	0.001	0.001	0.040	0.83	0.408	[− 0.001, 0.002]
Concomitant injuries	−0.085	0.028	−0.199	−3.00	0.003	[− 0.140, − 0.029]

Discussions

This study provides valuable insight into the HRQoL outcomes of adult patients with TBI in three low-to middle-income countries. By analyzing EQ-5D-5L index scores and EQ-VAS ratings alongside demographic and clinical variables, important patterns in acute post-TBI recovery were identified. The findings highlight both expected trends and unexpected outliers, showing the complexity of using the GCS as a predictor of HRQoL outcomes.

TBI registry

Our results showed that most TBIs occurred in trends similar to other countries, where males suffer more TBIs compared to females^{1,12,13,20,21}, mostly from RTI and due to falls^10,22. Despite the limited epidemiological data in our three studied LMICs, the patterns observed in our study are largely consistent with those reported in HICs. These findings suggest that certain epidemiological features of TBI may be broadly consistent across different economic and healthcare contexts, thus offering the opportunity for cross-national comparisons. Nonetheless, differences in healthcare access, availability of rehabilitation services, and long-term follow-up capacities still influence outcomes and should be further considered.

HRQoL measures - EQ-5D-5L, index scores and EQ-VAS

Our results showed that a significant percentage of adults with TBI experienced at least one HRQoL symptom at discharge. In most middle-income countries, follow-up and rehabilitation services are scarce, and most patients may have little healthcare guidance as they live with some symptoms post-discharge^1,8,12,21. Moreover, other studies show that traumatic injuries still pose difficulties in coping with activities that impact the quality of life, such as self-care activities, usual day-to-day activities, pain/discomfort, and anxiety or depression^23,24, dimensions on which our respondents ranked higher in reporting problems. This assumption can be further sustained by our results showing a weak association between GCS score and HRQoL, mostly due to having a majority of mild cases, similar to what other studies have found^25,26. However, these findings need to be further explored by comparing them to samples of patients with more severe TBIs to have better and more comprehensive data. Our multivariable regression analysis provided additional insight into the predictors of HRQoL. After adjusting for demographics and clinical characteristics, age, country, and concomitant injuries emerged as significant independent predictors of EQ-5D index, whereas sex, GCS, and length of stay were not. These findings suggest that broader demographic and injury context factors may play a greater role in shaping early HRQoL outcomes than initial neurological severity alone. The lack of an independent association between GCS and HRQoL in the adjusted model may reflect the influence of cultural perceptions, reporting differences across countries, and the disproportionate impact of concomitant injuries such as extremity trauma on self-reported quality of life^25,27,28.

A key finding in this study is the negative correlation observed between GCS scores and EQ-5D-5L index scores. While a higher GCS score typically indicates a better neurological status²⁶. Our results suggest that patients with mild TBIs reported lower quality of life scores, possibly due to post-concussion symptoms, cognitive impairments, or psychological distress. These results suggest that GCS alone is insufficient to predict long-term recovery, as other factors such as access to rehabilitation, psychological resilience, and social support system may play a crucial role in post-TBI quality of life²⁶. Patients with higher GCS scores, who are likely to have milder TBIs, might have higher expectations for their recovery and overall quality of life. Consequently, even slight impairments or residual symptoms post-injury could be perceived more negatively, leading to lower self-reported HRQoL scores on the EQ-5D-5L and VAS. This could be particularly true if patients compare their current health status to their pre-injury baseline, where they were presumably in good health^29,30. This can be aligned with the disability paradox, which suggests that individuals with significant disabilities often report a good quality of life, despite their functional limitations. In contrast, those with milder impairments may experience greater dissatisfaction due to the discrepancy between their expectation and their post-injury situation^31–33. On the other hand, patients with more severe TBIs (lower GCS scores) might perceive any level of recovery as a significant improvement, leading to relatively higher HRQoL scores despite their poorer initial condition^34–36.

Another explanation of our results might be the role of health literacy and patients’ psychological resilience. Patients with higher health literacy might be more aware of the long-term implications of TBI and thus more critical of their current health status. In contrast, those with lower health literacy might not fully grasp the potential complications and could report a better-perceived quality of life due to a lack of awareness of their condition’s severity^37,38. Psychological resilience could also play a role, where individuals with a stronger ability to cope with adversity may report a better quality of life, even when facing significant health challenges. Moreover, cultural differences in the perception of health and recovery could influence how patients rate their quality of life. In some cultures, there may be a tendency to underreport problems or to express a stoic attitude toward health issues, which could skew the EQ-5D-5L index scores³⁸. This cultural context could be particularly relevant in the countries studied, where societal norms might dictate a more reserved or positive reporting of health status, irrespective of actual physical or cognitive impairment. Our study did not account for other potential determinants of HRQoL, such as cultural norms, health literacy, or social support; however, our findings indicate the possibility of incorporating these factors in future research on TBI and quality of life.

Another consideration is the potential clustering of data points, particularly with a majority of patients having mild TBIs. The limited variability in GCS scores might have contributed to a less distinct correlation with HRQoL measures, as subtle differences in cognitive function could disproportionately affect the perceived quality of life. Additionally, the presence of comorbid conditions or social support systems could significantly impact patients’ perceived quality of life, further complicating the relationship between GCS and HRQoL. The right-skewed distribution for the index score indicates that while many patients recover to a high quality of life, there is still a need to identify and support those at risk of poorer outcomes. Future research should focus on the factors contributing to both high and low quality-of-life scores to improve targeted interventions for all TBI patients. The observed cross-country variation in EQ-VAS scores may partly reflect differences in the organization of TBI care and rehabilitation services across Armenia, Moldova, and Georgia. While all three countries face challenges like middle-income health systems, the capacity for pre-hospital care, access to neurosurgical units, and availability of rehabilitation services vary.

To our knowledge, this is the first study to explore the HRQoL for patients with a TBI in the Republic of Armenia, Georgia, and the Republic of Moldova. The study results are important, as health and quality of life measures are essential to assess health outcomes from the patient’s perspective to offer better rehabilitation and prevention programs and services, as well as a better allocation of resources to prevent and treat TBI. The strengths of the present study consist in the fact that it used a standardized TBI registry and a validated HRQoL tool in all three countries to collect the data in a systematic manner, so it can be comparable to other settings and contexts.

Limitations

Several limitations should be acknowledged. First, data were collected over a limited six-month period, therefore, causality between HRQoL, TBI-related symptoms, and long-term outcomes cannot be inferred. Linkage with primary care data over longer periods could strengthen future analyses. Second, the registry was designed primarily for documenting TBI diagnosis and outcomes, and did not capture other determinants (such as cultural factors, social support, or health literacy), which can influence self-reported health outcomes and recovery^39,40. Third, due to the lack of national EQ-5D-5L value sets in Armenia, Georgia, and Moldova, the Romanian value set was applied as a proxy. While methodologically acceptable, this may limit accuracy given cultural, economic, and healthcare system differences. Potential selection bias in EQ-5D-5L represents a potential limitation. Approximately one-quarter of patients (25%) did not complete the EQ-5D-5L and non-respondents were disproportionately those with severe TBI. This underrepresentation of more severely impaired patients may lead to an overestimation of HRQoL in the sample. Although sensitivity analysis suggested that the findings were robust, the potential for residual bias cannot be entirely excluded. Finally, the EQ-5D-5L is a generic rather than TBI-specific tool, which may not fully capture the consequences of brain injury. Future research could enhance comprehensiveness by combining registry data with TBI-specific HRQoL instruments.

Despite its limitations, this study provides valuable insights into the HRQoL of TBI patients based on data from a TBI registry implemented in different healthcare settings. Collecting comparable data from three countries over six months demonstrates how TBI registries can be implemented and expanded for longer periods. This approach supports data linkage with existing patient records and promotes longitudinal studies to identify recovery predictors, especially for moderate and severe TBI cases. Improving prevention and rehabilitation programs and services are crucial for better TBI recovery outcomes.

Conclusion

The study offers valuable data on the HRQoL of TBI patients from three countries, highlighting encouraging outcomes and areas needing further attention. The unexpected negative correlation between GCS score and index score draws attention to the complexity of assessing recovery and quality of life in TBI patients. A striking outcome of this study is the negative correlation observed between the GCS and HRQoL measures. Our findings highlight the variability in recovery paths, emphasizing that GCS alone does not fully predict long-term quality of life. While GCS remains a gold standard of acute TBI assessment, our findings may point to subtle post-injury recovery trajectories. Even patients with milder TBIs may continue to experience persistent symptoms that negatively affect their perceived health. Moreover, the presence of outliers in both mild and severe TBI groups reinforces and highlights the need for a multidimensional approach in post-TBI assessment to ensure robust capture of outcomes. Regression analysis further showed that age, country, and concomitant injuries were independent predictors of HRQoL, while sex, GCS, and length of stay were not, underscoring the influence of demographic and injury context factors. Simultaneously, cultural norms and different healthcare infrastructures across the Republic of Armenia, Georgia, and the Republic of Moldova influence long-term outcomes, thus highlighting the need to invest in consistent rehabilitation services and comprehensive patient education. By recognizing that injury severity alone is insufficient for predicting long-term well-being, policymakers and clinicians can tailor better interventions to the complex biological, psychological, and social factors that shape recovery.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

D.D., M.A.C., and C.P.A., contributed to the study’s conceptualization, methodology design, and supervision of data analysis. Material preparation and data collection were performed by A.T., N.C., and S.C. The first draft of the manuscript was written by P.M.M., D.D., M.A.C., and S.S. and reviewed by A.T., N.C., and S.C. C.P.A., P.M.M., D.D., M.A.C., and S.S. contributed to the interpretation of results. All authors commented on previous versions of the manuscript. All authors read and approved of the final manuscript and agree to be accountable for all aspects of the work.

Funding

This work was funded by the NIH-Fogarty International Center project iCREATE: Increasing Capacity for Research in Eastern Europe (National Institutes of Health, Fogarty International Center 2D43TW007261, Peek-Asa and Dulf, Co-PIs) and NIH-National Institute for Neurological Disease and Stroke project INITIATE: International Collaboration to Increase Traumatic Brain Injury in Europe (5R21NS098850, Peek-Asa and Dulf, Co-PIs).

Data availability

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.