Emotional, behavioural, and psychiatric problems are well documented in children with CHD and commonly persist after surgical correction,Reference Miatton, De Wolf, Francois, Thiery and Vingerhoets1 but there is scarce published data from low- and middle-income countries. Because surgeries in high-income countries are performed in a timely fashion, quality-of-life studies from high-income countries include mainly patients with corrected or palliated defects or those with mild lesions that do not require corrective surgery.
Separate quality-of-life data from healthcare-restricted settings in low- and middle-income countries are needed for a number of reasons, most significantly the much higher prevalence of uncorrected defects. Other key differences in these settings include delayed diagnoses, longer travel times to get to clinics and differing cultural expectations, for example, that children will engage in manual labour in support of the family income.Reference Iyer, Moreno and Fernando Caneo2,3 To our knowledge, a recent cross-sectional study in South India is the only large quality-of-life study in children with uncorrected cardiac defects.Reference Raj, Suhakar and Roy4
Emotional functioning is a key component of general quality-of-life with poor emotional functioning correlating with psychiatric morbidity in particular anxiety and mood disorders.Reference Bastiaansen, Koot, Ferdinand and Verhulst5
Our study is among the first to examine emotional quality-of-life in patients with uncorrected paediatric congenital and acquired heart disease in healthcare-restricted settings in low- and middle-income countries.
Over the course of thirty-one short-term surgical missions, we conducted a retrospective cross-sectional study using an objective, validated tool to measure emotional quality-of-life and to identify factors independently associated with poorer emotional quality-of-life in patients with uncorrected defects. We hypothesised that emotional quality-of-life would be independently associated with proxy measures of socio-economic hardship, symptom burden, and healthcare burden. We also collected data on patient-reported limitations in sports participation and hypothesised that this would be independently associated with proxy measures of symptom burden.
Materials and methods
Patient inclusion
Patients included in the study were children aged 6–17 years scheduled for corrective cardiac surgery during CardioStart International free compassionate cardiac surgery missions. The study period was September 2008 to November 2017 inclusive and included 31 surgical missions in 12 countries. Surgeries were in government, religious not-for-profit or private hospitals, where both surgery and peri-operative care were provided free-of-charge. The local medical/surgical team generated a shortlist of candidates on the basis of clinical indication, the need for specific surgical expertise of the visiting team, availability of surgical equipment, time constraints, and broader social factors that were likely to affect prognosis if a lesion was left untreated. The shortlist was reviewed and finalised on arrival of the visiting surgical team. We included children who were worked up for surgery but did not have surgery during the mission because of time constraints, equipment limitations, or other reasons.
Data collection
We administered the PedsQL TM 4.0 Generic Core scaleReference Varni and Seid6 which uses a Likert scale response separation of “never”, “almost never”, “sometimes”, “often”, or “almost always” to generate a score between 0 and 4. These scores were reversed and linearly transformed to scores of 0, 25, 50, 75, or 100 for each question. We reported on the five questions in the emotional functioning domain relating to fear, sadness, anger, trouble sleeping, and concern for the future. A total score for the emotional functioning domain was calculated for each patient with higher scores signifying better emotional quality-of-life. Scores were included only if four or more questions had been answered.
Data were supplemented by responses to a short questionnaire designed in collaboration with paediatric, cardiology, cardiothoracic surgery, and biostatistics faculty members from Oregon Health Sciences University and the University of Minnesota. This included a question on limitation in sports participation which used the same response separation detailed above. Further questions related to limitations in participation in activities for personal enjoyment and in activities with others and experiences of bullying.
In-person interviews with the patient were conducted by our data collator pre-operatively during the same admission in which the surgery was intended to be performed. During some of the missions, a translated copy of the questionnaire was available, while, in others, only verbal translation was available and was facilitated by a local translator. The patient was the subject of the questionnaire; however, in some cases, the parent/proxy was also present.
Demographic, social, and clinical data were supplemented by inspection of the patient clinical record. Data were recorded into a confidential REDCAP University of Minnesota database7 on the day of the interview where possible, and in all cases by the end of the surgical mission.
Statistical methods
Emotional functioning scores were reported as means with standard deviations and confidence intervals. Univariate association of demographic, social, and clinical factors with emotional functioning score and with patient-reported limitation in sports participation was analysed with Chi-Square, Fisher-exact, Student t-test, Wilcoxon rank sum, and Spearman’s rank correlation where appropriate. We used a preliminary threshold for significance of p < 0.05 and applied a Bonferroni correction to correct for multiple comparisons. Independent variables included in the regression analyses were age, gender, and any other factors which had p < 0.10 on univariate analysis. Variables which displayed collinearity or homoskedasticity on visual inspection by scatter plot were rejected from the model. Factors independently associated with emotional functioning scores were identified using multiple linear regression and factors independently associated with patient-reported limitation in sports participation were identified using ordinal regression. All analyses were performed in Stata 12.
Ethics
This study followed the requirements of the University of Minnesota (UMN-IRB no: 1307-M39481) and was subject to internal audit following each mission. Written permissions were obtained from local hospital administrations, and written consents were obtained from individual patients and/or their parent/guardian. Yearly reports of compliance were submitted to the University of Minnesota. There were no protocol deviations.
Results
One hundred and two patients were included (Table 1) from 12 countries (Fig 1): 2 low-, 6 low-middle, and 4 higher-middle-income countries (Fig 2).8 Atrial septal defect, ventricular septal defect, and patent arterial duct were the most common lesions (Fig 3). Half of the families reported a 50% loss in total household income since diagnosis and over 25% reported hunger in the household at least once a month (Table 1).
a p-Value < 0.002 threshold for significance following correction for multiple tests
b Numbers appearing in parenthesis in the clinical and demographic factors column are the number of patients in each age group for which this data point, i.e., lesion type was available – the first number is the age 6–10 group and the second is the age 11–17 group.
Ninety-four emotional functioning scores were included following eight exclusions (for answering less than four of the five questions). Mean emotional functioning score was 71.94 (SD 25.32) [95% CI 66.75–77.13] with no difference between children aged 6–10 years and children aged 11–17 years on specific emotions except for “concern for the future” for which older children reported lower scores (p = 0.033). On univariate analysis, acquired lesions, missing >5 weeks of school, and subjective limitation in walking up a flight of 10 steps and lifting an item equivalent to a 4 lb load all trended towards lower emotional functioning scores but none were significant following correction for multiple tests (Supplementary Table S1).
Eight factors were included in the regression model. Factors independently associated with lower emotional functioning score were a history of a parent with chronic disease or who had died, the patient having less than three siblings, and subjective limitation lifting an item equivalent to a 4 lb load (Table 2).
a Adjusted R2 = 0.26 from 46 observations
b p-Value < 0.05 threshold for significance
Eighty-one patients responded to the question on limitation in sports participation. On univariate analysis, patient-reported limitation at least “sometimes” showed a trend toward greater prevalence in older patients (Supplementary Table S2) which was not significant following correction for multiple tests. Six factors were included in the regression model. Experiencing symptoms at least once a month was the only independently associated factor (Table 3).
a Pseudo R2 = 0.20 from 51 observations
b p-Value < 0.05 threshold for significance
b p-Value < 0.05 threshold for significance
Sadness was reported at least “sometimes” by 51%, and bullying was reported at least “sometimes” by 22%. Sleep problems at least “sometimes” were associated with acquired lesions (p < 0.001), and limitation in activities for personal enjoyment at least “sometimes” showed a trend towards greater prevalence in female gender (p = 0.042) which was not significant following correction for multiple tests.
Discussion
Patient characteristics
Socio-economic hardship and lack of basic amenities available in high-income countries were common across all of our settings. Many of our patients experienced substantial loss of income since diagnosis likely mediated by long periods of school absence precluding ongoing parental employment, costs of medical care, and associated travel and medications. It must be stressed that without the surgical mission, it is unlikely that many of our patients would have undergone corrective surgery and equally unlikely that their quality-of-life living with an uncorrected lesion could be studied.
The vast majority of our patients had mild or moderate complexity congenital lesions.Reference Warnes, Williams and Bashore9 A small minority of patients had acquired lesions, mainly mitral valve rheumatic lesions, while only one patient had single ventricle physiology. In our settings, most patients born with complex lesions die prior to surgical workup often due to late presentation.Reference Mocumbi, Lameira, Yaksh, Paul, Ferreira and Sidi10 Anecdotally, some patients with complex lesions are transferred from remote healthcare-restricted settings to tertiary hospitals in the same country or even overseas surgical centres for self-funded or more rarely compassionate surgery.Reference Ramnarayan, Intikhab, Spenceley, Iliopoulos, Duff and Millar11
Comparison to previous studies
The cross-sectional study by Raj et al in South India is to our knowledge the only large quality-of-life study in children with uncorrected cardiac defects.Reference Raj, Suhakar and Roy4 Despite more complex lesions in their cohort, mean emotional functioning scores were similar to our own.
Other previous studies included only patients with corrected/palliated defects or those with mild lesions that did not require corrective surgery (Table 4), and direct comparison is therefore challenging. This said, our mean emotional functioning scores were similar to those of patients with corrected defects in Pakistan,Reference Ladak, Hasan, Gullick, Awais, Abdullah and Gallagher18 higher than those of palliated single ventricle patients in two high-income countries,Reference Eagleson, Justo, Ware, Johnson and Boyle15,Reference Mellion, Uzark and Cassedy16 and lower than those of healthy control groups in both low-middleReference Raj, Suhakar and Roy4 and high-income countries.Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni12,Reference Kwon, Mussatto, Simpson, Brosig, Nugent and Samyn14,Reference Mellion, Uzark and Cassedy16,Reference Uzark, Zak and Shrader17
a Warnes CA, Liberthson R, Danielson GK, et al Bethesda Conference Report Task force 1: the changing profile of congenital heart disease in adult life. J Am Coll Cardiol 2001; 37: 1170–1175.
Established predictors of poorer emotional quality-of-life in patients with corrected or palliated defects or those with mild lesions that did not require corrective surgery include: low family income,Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni12,Reference Uzark, Zak and Shrader17,Reference Drakouli, Petsios, Giannakopoulou, Patiraki, Voutoufianaki and Matziou19–Reference Xiang, Su and Liu22 short duration of parental education,Reference Niemitz, Gunst and Hovels-Gurich23 single parent status,Reference Goldbeck and Melches24 ethnic minority status,Reference Wang25 absence of spirituality,Reference Wang25 at least one regular symptom,Reference Drakouli, Petsios, Giannakopoulou, Patiraki, Voutoufianaki and Matziou19,Reference Bertoletti, Marx, Hattge and Pellanda26 subjective exercise limitation,Reference Marino, Cassedy and Brown27 NYHA class,Reference Niemitz, Gunst and Hovels-Gurich23 ROSS Class,Reference Ong, Teh, Darshinee, Omar and Ang28 objective exercise capacity,Reference Drakouli, Petsios, Giannakopoulou, Patiraki, Voutoufianaki and Matziou19,Reference Dulfer, Helbing, Duppen and Utens29 specific lesion types,Reference Berkes, Varni, Pataki, Kardos, Kemeny and Mogyorosy13,Reference Ladak, Hasan, Gullick, Awais, Abdullah and Gallagher18,Reference Idorn, Jensen, Juul and Overgaard30–Reference Kahr, Radke, Orwat, Baumgartner and Diller32 lesion complexity,Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni12,Reference Berkes, Varni, Pataki, Kardos, Kemeny and Mogyorosy13,Reference Mellion, Uzark and Cassedy16,Reference Ernst, Marino and Cassedy33–Reference Landolt, Valsangiacomo Buechel and Latal37 previous surgeries,Reference Svensson, Idvall, Nilsson and Liuba38 number of clinic appointments,Reference Marino, Cassedy and Brown27,Reference Ernst, Marino and Cassedy33 school absence,Reference Niemitz, Gunst and Hovels-Gurich23,Reference Knowles, Day, Wade, Bull, Wren and Dezateux39 and medication burden.Reference Ladak, Hasan, Gullick, Awais, Abdullah and Gallagher18,Reference Marino, Cassedy and Brown27,Reference Ernst, Marino and Cassedy33,Reference Knowles, Day, Wade, Bull, Wren and Dezateux39
Our trend towards poorer emotional quality-of-life in adolescents (though not statistically significant) was in contrast to the findings of Raj et al who found that 8–12 year-olds report poorest emotional quality-of-life relative to controls.Reference Raj, Suhakar and Roy4 It was also in contrast to studies in patients with corrected lesions which found younger children to report poorest emotional quality-of-life.Reference Berkes, Varni, Pataki, Kardos, Kemeny and Mogyorosy13,Reference Mellion, Uzark and Cassedy16,Reference Idorn, Jensen, Juul and Overgaard30,Reference Manlhiot, Knezevich, Radojewski, Cullen-Dean, Williams and McCrindle40 Our findings may relate to the societal role of adolescents in many of our settings where there is often an expectation to engage in manual labour to contribute to household income or to support a family business, though we did not formally record rates of this practice.
Our finding that having three or more siblings was independently associated with better emotional quality-of-life might be explained by a pseudo-parental role played by older siblings as opposed to younger siblings who might compete for parental attention. Prior studies in both CHDReference Manlhiot, Knezevich, Radojewski, Cullen-Dean, Williams and McCrindle40,Reference Im, Yun and Lee41 and other childhood chronic diseasesReference Inati, Sleiman, Akiki, Lteif, Kahale and Abbas42,Reference Ozyazicioglu, Avdal and Saglam43 are divided on the effect of siblings on emotional quality-of-life.
Our finding that a history of parental chronic morbidity or death was associated with poorer emotional quality-of-life is difficult to interpret in the absence of a control arm, although it is conceivable that past experience of parental illness may inform a child’s experience of their own disease.
Multiple previous studies have demonstrated that girls with heart disease report lower emotional functioning scores than boys,Reference Ladak, Hasan, Gullick, Awais, Abdullah and Gallagher18,Reference Ernst, Marino and Cassedy33 but we were unable to replicate this finding. Girls with heart disease are less likely to display disruptive behaviours than boys,Reference Spijkeboer, Utens, Bogers, Verhulst and Helbing31 and therefore problems that do arise may present late or be missed entirely. There may also be a cultural tendency in some settings to be more protective of girls resulting in overly cautious activity restriction.Reference Ladak, Hasan, Gullick, Awais, Abdullah and Gallagher18
Aside from corrective surgery, regular exercise is the only evidence-based intervention to improve emotional quality-of-life.Reference Dean, Gillespie and Greene44 Our rates of patient-reported limitation in sports participation were similar to those of a large cohort in the US.Reference Dean, Gillespie and Greene44 Though we found that patient-reported limitation in sports participation was independently associated with symptom burden, we suspect that some of our patients reported limitation because of externally imposed restriction. While previous studies in high-income countries have shown that restriction is in most cases doctor-initiated, in a minority it is parent- or school-initiated and often inappropriate.Reference Dean, Gillespie and Greene44 We feel that inappropriate restriction is a particular risk in our settings as lack of regular clinic appointments and access to evidence-based exercise recommendations may result in patients following advice that is no longer age-appropriate, is parent- or school-initiated, and is often overly cautious.
Study limitations
Our study was retrospective, had small numbers, and was cross-sectional. Longitudinal data including data on long-term post-operative emotional quality-of-life would have been highly challenging to collect as our patients commonly changed their address and phone number prior to 1-year follow-up. The lack of sibling or age/sex-matched control arm and the absence of any available data in the literature on population norms in our settings was also a limitation.
Because we conducted the interview simultaneously with other components of the pre-operative workup, in some cases both patient and parent/proxy were present and parent/proxy input may have influenced patient scoresReference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni12,Reference Mellion, Uzark and Cassedy16,Reference Latal, Helfricht, Fischer, Bauersfeld and Landolt20,Reference Ong, Teh, Darshinee, Omar and Ang28 particularly in the adolescent age group.Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni12 In some settings, use of more than one local language/dialect presented translation difficulties which could have influenced interpretation of questions.Reference Ladak, Hasan, Gullick and Gallagher45
Missing data affected clinical variables in particular which decreased our numbers and introduced a potential sampling bias. Low numbers for clinical variables also increased the risk of type II error in the univariate analysis, and we suspect that with larger numbers we would have found lower emotional functioning scores in patients with severe disease.
Though previous studies have shown that socio-economic factors accounted for two-thirds of the variance in quality-of-life scores,Reference Ernst, Marino and Cassedy33 the exclusion of clinical variables from the regression analyses may have affected model accuracy.
Future directions in practice
Short-term surgical missions offer the possibility for novel one-off interventions that may improve emotional quality-of-life. Missions could provide plain-language pamphlets translated into local dialects for families to explain their condition to schools and combat stigma around exercise restriction and surgical scars. Informational leaflets for primary care doctors or community nurses might serve to minimise anxiety-provoking referrals to emergency departments. Social events during the surgical mission could connect families and encourage formation of local support groups. Didactic teaching to local cardiologists and nurses could include modules on up-to-date exercise recommendations.
One-off parental workshops with components of problem prevention therapy, psychoeducation, and parenting skills have been shown to improve maternal mental health and decrease number of days missed from school in CHDReference McCusker, Doherty and Molloy46 and could be replicated in these settings with the aid of local translators. Supervised exercise testing can be used therapeutically to demonstrate to patient and parent that vigorous exercise can be performed successfully and safely thus combating exercise aversion.Reference McCusker, Doherty and Molloy46
Long-term psychotherapy may be required but access to child psychology and mental health services varies greatly between settings. Nepal has only one outpatient child and adolescent mental health clinic, one trained adolescent psychiatrist, no post-graduate specialist training programme and few clinical psychologists specialising in paediatrics.Reference Chaulagain, Kunwar, Watts, Guerrero and Skokauskas47 In contrast, Peru is currently undergoing wide-scale expansion of community-based child and adolescent mental health services in remote areas.Reference Toyama, Castillo and Galea48
Future directions in research
Models for conducting research during surgical missions need to be refined.Reference Murala, Karl and Pezzella49 Collaboration between different charitable organisations would increase patient numbers and potentially answer key questions, and data collected could be incorporated into existing international registries.50 Primary screening programmes could be used to recruit local controls greatly improving study quality.
Longitudinal projects require “buy-in” from the local team which can be a challenge. Local ownership over research projects could be improved by: (a) establishing a single local contact for the research component of a mission as distinct from the clinical component, (b) one-on-one delivery of educational modules during the mission teaching basic descriptive research methodology, (c) short-term research fellowships for the local team member to work on the project at an overseas institution, (d) being listed as an author on a poster or journal article.
Future studies using specialised psychiatric screening tools could usefully define the burden of psychiatric morbidity in these settings and objective exercise testing could discern between externally imposed exercise restriction versus true limitation.
Conclusions
Our study is among the first to describe emotional quality-of-life in uncorrected paediatric congenital and acquired heart disease in healthcare-restricted settings in low- and middle-income countries. We found that some of the factors associated with better emotional quality-of-life were similar to those identified in previous studies in patients with corrected defects. Given the scarcity of formally trained psychologists in many of our settings the role of early screening for emotional, behavioural, and neuropsychiatric problems falls to local cardiologists, surgeons, and nurses. Age- and disease-appropriate exercise recommendations are especially important in these settings as exercise is a low-cost, evidence-based intervention. In addition to corrective surgery and exercise, numerous other interventions which are practicable during surgical missions might improve patient emotional quality-of-life.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1047951120000220
Acknowledgements
No specific acknowledgements
Financial Support
This research was supported by the National Institute of Heath’s National Center for advancing Translational Science, grant ULTR002494. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Health’s National Center for Advancing Translational Science.
Conflicts of Interest
None to disclose