This study was reviewed and approved by the Food, Medicine and Healthcare Administration and Control Authority of Ethiopia, the National Health Research Ethics Review Committee of the Ethiopian Ministry of Science and Technology, Amhara Regional Health Bureau Research Ethics Review Board Committee, the London School of Hygiene and Tropical Medicine (LSHTM) Ethics Committee, and Emory University Institutional Review Board. Written informed consent in Amharic was obtained prior to enrolment from participants. If the participant was unable to read and write, the information sheet and consent form were read to them and their consent recorded by thumbprint.

This case-control study was nested within a clinical trial of two alternative surgical treatments for trichiasis. For the trial 1000 trichiasis cases were recruited, and these were also enrolled into this QoL study. Cases were defined as individuals with one or more eyelashes touching the eyeball or with evidence of epilation in either or both eyes in association with tarsal conjunctival scarring. People with trichiasis from other causes, recurrent trichiasis and those <18 years of age were excluded. Trichiasis cases were identified mainly through community-based screening. Trichiasis screeners and counsellors (Eye Ambassadors) visited every household in their target village, identified and referred trichiasis cases to health facilities where surgical services were provided. Some cases self-presented or were referred by local health workers. Recruitment was done mainly from three districts of West Gojam Zone, Amhara Region, Ethiopia between February and May 2014. This area has one of the highest burdens of trachoma worldwide [17].

We recruited 200 matched controls to every fifth consecutive trichiasis case. Controls were individuals without clinical evidence or a history of trichiasis (including epilation), and who came from households without a family member with trichiasis or a history of trichiasis. Controls were individually matched with every fifth trichiasis case by location, sex and age (+/- two years). The research team visited the sub-village (30–50 households) of the trichiasis case that required a matched control. A list of all potentially eligible people living in the sub-village of the case was compiled with the help of the sub-village administrator. One person was randomly selected from this list using a lottery method, given details of the study and invited to participate if eligible. If a selected individual refused or was ineligible, another was randomly selected from the list. When eligible controls were not identified within the sub-village of the index case, recruitment was done in the nearest neighbouring sub-village, using the same procedures.

The VRQoL and HRQoL were administered orally by six trained Amharic speaking interviewers, because of the low literacy rate amongst participants. Data from trichiasis cases were collected at health facilities at the time of enrolment into the clinical trial, prior to surgery. Data from the controls were collected at their homes.

Data were also collected on general health problems and self-rated socioeconomic status (SES). For the self-rated socio-economic status, participants were asked to rate the wealth of their households in relation to other households in their village by choosing one of the following options: (1) very poor, (2) poor, (3) average, (4) wealthy or (5) very wealthy. In addition, data were collected on social relations, marriage and sleeping, through semi-structured questions including: “Do you feel ashamed or embarrassed due to the trichiasis?”, “Do you worry that you may lose your remaining eyesight due to the trichiasis?”, “Do you have a sleeping problem?” and “If yes, do you think your sleeping problem is related with the trichiasis?”

Presenting LogMAR (Logarithm of the Minimum Angle of Resolution) visual acuity at two metres was measured using “PeekAcuity” software on a Smartphone in a dark room [21]. We assessed contrast sensitivity with a prototype smartphone based test that presents the individual with calibrated grey scale spots against a white background, which they have to identify by touching the screen (www.peekvision.org). Unlike visual acuity, which is measured with high contrast, contrast sensitivity perhaps more accurately reflects the person’s everyday visual experience in varying conditions. Patients with normal visual acuity may have profound contrast sensitivity impairment. Therefore it is useful to measure contrast sensitivity while investigating VRQoL, as impairment could lead to decreased functioning and quality of life [22]. The ophthalmic examination was conducted using a 2.5x binocular magnifying loupe and a bright torch. Clinical signs were graded using the Detailed WHO FPC Grading System [23].

The sample size was calculated with the aim of detecting a three point difference in mean QoL score between trichiasis cases and controls [5]. The sample size of 1000 cases and 200 controls has 90% power to detect even minimal effect of trichiasis on QoL with an effect size of about 0.27 (effect size = QoL score difference (3)/SD (11)) with a Type I error of 5%.

Data were double-entered into Access (Microsoft), cleaned in Epidata 3.1 and transferred to Stata 11 (StataCorp) for analysis. Data were analysed as follows:

Cases and controls were adequately matched for age (Table 1) and for geographical distribution across the 20 administrative units where recruitment was conducted. There were proportionately slightly more females among the 200 controls (83%) than the 1000 cases (76%). This occurred by chance as the 200 trichiasis cases used to determine the control matching characteristics had proportionately more females than in the full group of 1000 trichiasis cases. The trichiasis cases were more likely to be widowed or divorced, be from poorer households, and report other health problem in the past month than controls. For the analysis of socio-economic data we combined the “very poor” with the “poor” group and the “very wealthy” with “wealthy” group because of small numbers at both ends of the SES distribution. Trichiasis cases had substantially lower visual acuity scores (median LogMAR, 0.35; IQR, 0.15 to 0.6) than the controls (median LogMAR, -0.05; IQR, -0.1 to 0.1; Wilcoxon ranksum test, p<0.0001). Among trichiasis cases vision in the better eye was 6/18 or better in 63%, and 50% had minor trichiasis. The median duration of trichiasis among the cases was 5 years (IQR, 2–10).

The trichiasis cases had substantially lower VRQoL scores than the controls in all four subscales (p<0.0001) (Table 2). The largest differences between cases and controls were found for visual symptoms (mean difference, 51.5; 95%CI, 48.5–54.5) and the smallest in general functioning (mean difference, 24.1; 95%CI, 21.1–27.1). In a sub-group analysis, trichiasis cases with normal vision had significantly lower VRQoL in all subscales than controls with normal vision (Table 3).

The relationship between VRQoL and various demographic and clinical characteristics among individuals with trichiasis are presented in Table 4. VRQoL scores were lower in all domains with increasing age, severity and duration of trichiasis and with decreasing visual acuity and contrast sensitivity scores. Scores were also lower in females, illiterate individuals and in cases with central corneal opacity.

In multivariable analysis, lower overall VRQoL scores were found in those with longer trichiasis duration, central corneal opacity, visual impairment and poor contrast sensitivity (Table 4). In addition to these, older age and female gender were associated with lower VRQoL score in the general functioning subscale. In the model, people with Major Trichiasis had lower VRQoL score in the visual symptom domain after controlling for potential confounders (p = 0.011), however, this is confounded by duration of trichiasis and was therefore dropped from the final model (Table 4).

The trichiasis cases had substantially lower overall HRQoL (p<0.0001) and overall self rated health (p<0.0001) scores than the controls (Table 2). Strikingly, 55.4% of trichiasis cases rated their overall QoL “poor” or “very poor” compared to only 9.5% of controls (p<0.0001) and 5.6% of cases rated their overall QoL “good” or “very good” compared to 59.5% of controls (p<0.0001). Across all four domains there were substantial differences in the QoL scores of cases and controls (all p<0.0001, Table 2). The largest difference was seen in the physical health domain (mean difference, 32.4; 95%CI, and 30.1–34.7). Trichiasis cases with normal vision had significantly lower HRQoL in all subscales than controls with normal vision (Table 3).

Among trichiasis cases, overall HRQoL and the four domain scores decreased with increasing age (except for environment domain), decreasing self-rated wealth, visual acuity and contrast sensitivity scores (Table 5). They were also lower in divorced/widowed, illiterate individuals, females and those with other health problems in the past month. Daily labourers, the unemployed and those with central corneal opacity had lower overall QoL and domain scores except for the environment domain. Participants with longer duration trichiasis had lower physical, psychological and social domain scores.

Multivariable analyses identified predictors of HRQoL among trichiasis cases (Table 5). Lower self-rated wealth was associated with lower QoL scores in all domains. Poorer overall QoL was related to not having a marriage partner, visual impairment, being a daily labourer and presence of other health problems. Older participants, females, the unemployed, those with visual impairment, poor contrast sensitivity score and other health problems were associated with lower physical domain scores. Daily labouring, not having a marriage partner and presence of other health problems were associated with lower psychological domain scores.

Among the 200 controls, 198 (99%) reported no ocular pain or discomfort. In contrast, among the 1000 trichiasis cases, 143 (14.3%), 281 (28.1%) and 562 (56.2%) reported mild, moderate and severe ocular pain or discomfort, respectively (p<0.0001). The cases reported the following effects of trichiasis: 596 (59%) felt ashamed or embarrassed; 913 (91.3%) worried that they might lose their remaining eyesight; 70 (7.0%) had been troubled in their marriage and ignored by their marriage partner; 681 (68.1%) reported sleeping problems, largely due to pain (675/681 (67.5%)) from the trichiasis.

Satisfying the known-groups difference criteria, the trichiasis cases had significantly lower VRQoL and HRQoL scores in all domains (p<0.0001) than the controls (Table 2). With respect to convergence validity, worsening visual acuity and contrast sensitivity scores, trichiasis duration and central corneal opacity were significantly associated with lower scores in all four domains of VRQoL (Table 4). The VRQoL data were reliable after being assessed for internal consistency with a Cronbach’s alpha: coefficients of >0.80. The overall QoL data showed very high internal consistency with a Cronbach’s alpha of 0.90. The physical health, psychological, social and environment domains demonstrated internal consistency with Cronbach’s alpha of 0.87, 0.65, 0.47, and 0.64, respectively.

Overall, the VRQoL of trichiasis cases was substantially lower in all domains compared to controls. When we restricted the analysis to people with a visual acuity of 6/18 or better, the difference in VRQoL was of a similar magnitude and was highly significant. This is an important observation, which demonstrates that trichiasis reduces VRQoL even before impairment of visual acuity develops.

We found that, of the four sub-scales, the one with the largest difference was the visual symptom subscale. This is a composite of questions about visual functioning, pain/discomfort, glare and light/dark adaption. The general functioning subscale showed the smallest difference compared to controls. This may be because approximately two-thirds of trichiasis cases had normal vision, and this subscale includes items on vision difficulties and role limitation.

Consistent with our study, several other studies have demonstrated the effect trichiasis has on physical functioning [26–28]. A population-based study in Tanzania found trichiasis without visual impairment results in limitation of physical functioning in women that was comparable to limitation associated with visual impairment from other causes [27]. A qualitative study of 23 women with trichiasis in Niger (without a control group for comparison) reported trichiasis had marked effects on the general well-being of these individuals and was linked to physical disability and inability to work and earn an income [26]. In a study conducted in southern Ethiopia, 61% of trichiasis cases reported difficulty in physical functioning including walking, recognizing faces and performing day-to-day farming activities [28].

Among trichiasis cases, VRQoL was significantly lower with central corneal opacity, increasing trichiasis duration and decreasing visual acuity and contrast sensitivity scores. This suggests that trichiasis cases had significantly lower contrast sensitivity than the controls. Other studies have found conditions such as dry eyes and reduced tear break-up time, resulting from progressive conjunctival scarring and ocular epithelial tissues damage, are associated with reduced contrast sensitivity score [29,30]. Impaired contrast sensitivity score can greatly affect the person’s ability to recognise objects and perform daily activities under different conditions. A recent study conducted on glaucoma patients revealed that contrast sensitivity plays a major role in daily functioning and VRQoL; and the contrast sensitivity score was correlated with VRQoL indicators such as facial recognition, finding objects, motion detection and general vision [22]. Poor contrast sensitivity has been associated with physical injuries from accidents, suggesting that it would greatly hamper overall well being [31,32]. The association between trichiasis severity and the visual symptom subscale was weakened after including trichiasis duration in the multivariable regression model. Severity and duration are not independent of each other. The general functioning subscale of the VRQoL and physical health domain in HRQoL were lower in females than males. Although the reason is not apparent, a similar finding has been reported in a Tanzanian study: women without visual impairment were more likely to have functional limitation than their male counterparts [27].

Trichiasis cases had a poorer HRQoL than controls in all domains, again even without visual impairment. Strikingly, the physical domain in the WHOQOL-BREF, which includes questions on pain and discomfort, had the highest mean score difference between cases and controls, emphasising the suffering trichiasis causes. Among the QoL domains, the environmental domain had the lowest score in both trichiasis cases and controls. This domain is built from items such as satisfaction with financial resources, access to health service, transport, information and leisure activities, which generally have low availability in the communities where this study was conducted. Hence, participants would be anticipated to have a lower rate of satisfaction for these items.

The WHOQOL-BREF has been used to assess HRQoL of entropion patients (with and without trichiasis) [5]. Apart from the environment domain scores of the trichiasis cases, the average QoL scores of trichiasis cases and controls in all domains in the Indian study were generally lower than those we recorded in Ethiopia [5]. This difference could be attributed to two things. Firstly, perceptions towards QoL in Indian and Ethiopian communities could be different. Secondly, in the Indian study, all participants were recruited from hospital, compared to community-based recruitment in our study. Hospital participants might be more likely to report poor QoL than people in the community [14]. In contrast to trichiasis, there is an extensive literature about the impact of cataract on QoL [7,9–11,33]. However, there are fundamental differences between cataract and trichiasis in the nature of visual loss, pain and other symptoms they cause.

These tools have previously been reported to be valid and reliable in studies conducted in similar settings to this study [9–11,14,18–20]. In this study, both the VRQoL and HRQoL data measured what they were intended to measure (construct validity) by demonstrating significant differences in the scores between groups known to be different; cases and controls had lower and higher scores respectively. The VRQoL data also showed that sub-scales correlate well with measures of similar constructs (convergent validity) such as visual acuity and contras sensitivity where worsening in these measures is associated with lower VRQoL subscale scores. There was evidence of higher homogeneity among the items in each VRQoL subscale (internal consistency) than the generally accepted criteria of >0.70. In the HRQoL data, the overall QoL and the physical health domain items were internally consistent and reliable in measuring the same construct, while the psychological, social relations and environment domains had less internal consistency. Similar psychometric properties have been reported in the field trial results of this tool in other countries [14]. The lower alpha score for the social relationship domain is anticipated as its analysis is based on three items instead of the generally recommended minimum four for evaluating internal consistency [6]. Hence, this domain’s results should be interpreted in caution, as there was insufficient evidence that the items in this domain are always measuring the same construct.

This is a large case-control study examining the impact of trichiasis on QoL. We used tools validated in settings similar to this study setting. The VRQoL tool has been validated in Kenya while the HRQoL tool has been tested and used in other studies in Ethiopia. The study has some limitations. Perfect matching was not achieved in terms of gender, resulting in more females in the controls than the trichiasis cases. However, all comparisons between cases and controls were adjusted for gender, age and self-rated wealth. A community-based screening method was employed to identify trichiasis cases. Although we think that this was an efficient and comprehensive approach to finding cases, it is possible that some cases might have been missed, particularly those with mild disease who may be less likely to come forward, which could lead to an overestimation of the QoL scores for trichiasis cases in general. Trichiasis is strongly associated with a poorer QoL, However, the cross-sectional nature of this study precludes us from drawing definite conclusions about causality. This question is being investigated by reassessing this group of people one year after surgery.

In this Ethiopian population, we found that trichiasis cases have significantly lower VRQoL and HRQoL than controls regardless of visual impairment. The results provide solid data for advocacy and encourage programme leaders and funders to secure resources to promote trichiasis intervention. Trichiasis inflicts considerable physical and psychosocial trauma including sleep disturbance, low self-esteem and possibly a less stable marriage. The burden of trichiasis goes beyond visual loss. Timely treatment is needed not only to prevent visual loss but also alleviate physical and psychological suffering and social exclusion of trichiasis patients, thereby improving their physical and psychological health, general functioning and social relations. The comprehensive SAFE strategy is needed to prevent the development of trichiasis. The long-term effect of trichiasis surgery on VRQoL and HRQoL in trichiasis patients needs to be measured in longitudinal studies.