Introduction

Prematurity is characterized by physical and neurological immaturity of infants at birth, which can have negative impacts on several biological systems [1–3]. Consequently, individuals born preterm may experience recurring health problems, in addition to developmental delays or disorders at various points of development [4–6].

In recent decades, scientific research on infants who are born preterm no longer focuses only on specific diseases or isolated areas of development, such as motor skills, cognition, language, or personal/social behavior. Instead, it has assessed the impact of premature birth on multidimensional holistic outcomes, such as functional abilities and quality of life (QoL). In particular, QoL has become a very important outcome that should be considered in matters of public health and epidemiology, as it complements traditional information on mortality and morbidity [7, 8].

Despite the increasing number of studies assessing QoL in preterm individuals, these have applied a wide variety of methodological approaches, leaving several questions unanswered. Analyzing data from multiple studies across this area of research may provide answers to some of these questions. We previously conducted a systematic review [6] that aimed to provide a general survey of the development and QoL of children born preterm at preschool and school age. Our analysis indicated that children born prematurely are at an increased risk of delay in several areas of development. However, insufficient findings were observed with respect to QoL outcomes. This was because there was a limited range of ages included in the review. Samples were only available for children aged between 3 and 12 years. Furthermore, we did not perform a critical analysis of the methods used to assess QoL.

Recent systematic reviews that have exclusively investigated QoL in preterm infants [9, 10] have only included infants with very low birth weight [9] and did not analyze factors associated with QoL in this population. They also did not provide a critical analysis of assessment instruments or their applicability. Furthermore, the most recent reviews only searched databases up to 2007, thus justifying an update of findings in this area.

The purpose of the present study was to review the most recent literature regarding factors associated with QoL in preterm individuals. We aimed to critically analyze the concept of QoL assessment, focusing on the dimensions of this construct, instruments, and procedures for data collection and analysis. Our review was guided by the following questions: (1) What are the characteristics and psychometric properties of the instruments used to assess QoL in preterm individuals? (2) What methods of data collection and management are used for such instruments? (3) Where reported, is there any agreement between measures of self-reported QoL and parental measures of QoL? (4) What factors are associated with QoL in preterm individuals?

Methods

We adhered to Preferred Reporting Items for Systematic Reviews and Meta-Analyses [11] when searching for and selecting articles in the following databases: PubMed, PsycARTICLES, PsycINFO, SciELO, and LILACS. Our search terms were determined by searching Mesh (PubMed), Decs (SciELO and LILACS), and Index (PsycINFO and PsycARTICLES) terms and descriptors listed in previous relevant studies. In selecting our search terms, we attempted to achieve high sensitivity, at the cost of low specificity. The search term combinations were: (Infant, Premature OR Premature Birth OR Prematurity OR Preterm) AND (Quality of Life OR Health related quality of life OR Health-related quality of life OR Personal Satisfaction OR Well being OR Well-being OR Life satisfaction). The detailed search method for each database is presented in the supplementary material (Online Resource 1).

We included empirical studies with observational designs assessing QoL in individuals who were born preterm (gestational age [GA] <37 weeks), regardless of birth weight, from the perspective of the individuals themselves or their main caregiver (parents or legal guardians). We included all articles published between January 2007 and January 2015, in English, Portuguese, or Spanish.

Review articles, meta-analyses, commentaries, editorials, letters, and clinical trials were excluded. We also excluded studies that evaluated parents’ or caregivers’ QoL, studies reporting children’s QoL from the perspective of individuals other than parents or legal guardians, and studies evaluating QoL only according to clinical indicators or that were restricted to environmental factors.

Two authors selected the studies. Figure 1 illustrates that 375 articles were initially identified in our database search. After both authors systematically applied the inclusion and exclusion criteria, 22 articles remained.

Fig. 1
figure 1

Flowchart of article selection

Full size image

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement was used to analyze the quality of reporting in the selected studies [12]. The STROBE statement recommends topics that should be included in an accurate and complete report of an observational study. It provides a checklist of 22 items related to the title, abstract, introduction, methods, results, and discussion. Appropriate reporting of studies is important to assess their strengths and weaknesses, and the generalizability of findings.

Additionally, we analyzed the psychometric properties of the QoL instruments used in the studies. For this specific analysis, we conducted additional literature searches to identify articles about the psychometric parameters of the instruments.

We organized the samples of preterm participants according to GA, using categories proposed by the World Health Organization (WHO) to classify prematurity as follows: extremely preterm (<28 weeks), very preterm (28–<32 weeks), and moderately preterm (32–<37 weeks) [13].

Results

Study characteristics

The majority of studies (17 studies; 77 %) used a prospective longitudinal design [14–30]. These samples were recruited in the neonatal period. Only two studies conducted repeated QoL measures at 14, 19, and 28 years of age in the same cohort of individuals who were born preterm [21, 26]. Four studies implemented retrospective longitudinal designs [31–34]. One study conducted a cross-sectional design [35].

Longitudinal designs can identify possible predictors of QoL outcomes. However, there is an increased chance of attrition during follow-up (dropouts) in such studies. Sample sizes ranged from 43 to 630 preterm participants. In half of the studies, non-response bias should be considered, as there were non-response rates (dropouts) of more than 25 % of the target population. Nevertheless, the vast majority of these compared sociodemographic and perinatal clinical characteristics between respondents and non-respondents, and considered any differences when discussing their findings. Only one study did not report having taken such methodological care [35].

There were other strengths in the literature, including good representativeness, with both population-based and multicenter samples used in 11 studies (50 %). Fourteen studies (64 %) were analytical and included a control group consisting of full-term healthy individuals. All of the studies attempted to control for confounding variables through matching and/or during the analyses. The most commonly controlled variables were gender, age, parental level of education, and socioeconomic status. The only exception was the study by Schiariti et al. [31], which identified differences between the group of preterm children and the control group in terms of gender and family income. However, they did not control for these differences in their analyses, nor did they highlight the risk of bias this introduces in their discussion.

We observed great variety in the methodological approaches used to assess QoL in preterm individuals. This culminated in significant diversity in the ages of samples and the way in which the instruments were applied. Figure 2 illustrates that almost half of the studies assessed QoL in children under 12 years of age, and most of these chose parents as the only source of information about the children’s well-being.

Fig. 2
figure 2

Distribution of studies according to the source of information on quality of life assessments and age group (n = 22)

Full size image

As the age of participants increased, the source of information regarding QoL changed. All studies conducted with adolescents or adults relied on self-report for QoL assessment. Only two articles dealt with QoL in preterm individuals from both the parents’ perspective and that of the individuals themselves [17, 20].

Table 1 presents the instruments and measures used in the studies. Sixteen different instruments were used to assess QoL. They were mostly generic measures (i.e., designed to assess QoL in the general population). The only exception was the London Handicap Scale (LHS) [26], which is a condition-specific instrument designed to assess QoL in individuals with chronic diseases. Considering that the immaturity of preterm neonates may affect them in a multisystemic way, generic instruments are the most appropriate tools to assess QoL in this population [36].

Table 1 Instruments and measures used to assess QoL
Full size table

All of the self-report instruments followed the three basic characteristics specified by the WHO: subjectivity, multidimensionality, and influence of positive and negative dimensions [37]. However, subjectivity can be affected by instruments that assess QoL in children based only on the parents’ perspective (parental report) because only the parents’ perception and concerns about their children are considered.

Sixteen articles (73 %) assessed QoL using self-administered questionnaires. Eleven of the instruments (68 %) have reported excellent or good psychometric properties. Three instruments (Fragen zur Lebenszufriedenheit-Module, Visual Analog Scale, and LHS) have been subjected to preliminary psychometric assessment. Just one instrument (Child Health Questionnaire—Parent Report 28) should be considered cautiously as it has been shown to be insufficiently reliable when each domain is analyzed separately. Only two instruments have no published psychometric evaluations [14, 30]. Notably, in two studies the authors adapted instruments, which affected the accuracy and precision of their results [33, 35].

Regarding the quality of reporting in the studies, most (77 %) included more than 70 % of the STROBE items, which demonstrates satisfactory description quality. We consider six observational studies (27 %) to have moderate description quality as only between 50 and 70 % of STROBE items were included [21, 24, 28, 31, 32, 35]. Only two observational studies included less than half of the STROBE items, consequently failing to provide important methodological information or discussion of the findings [23, 34]. The STROBE items that were most frequently missing included information about missing data, the way in which missing data were considered in the analysis, and information on the number of participants with missing data for each variable of interest.

Quality of life in children born preterm

Table 2 presents a summary of the ten studies that assessed QoL in preterm infants aged 1 month to 12 years. Seven studies compared QoL assessed using parental report between preterm children and full-term children (control group), with samples aged between 2 and 11 years [14, 16–18, 31–33]. Worse QoL in the children was reported by parents of preterm children compared with parents of the control group. This difference remained even after adjusting for gender, age, and socioeconomic status [14, 18, 33]. In these studies, parents of preterm children at preschool or school age reported more problems in various QoL domains, primarily related to physical health, behavior, or functionality. They also reported a higher impact of those problems on their own lives [31].

Table 2 Characteristics of the studies and summary of main results related to QoL of children born preterm (1 month–12 years of age; n = 10)
Full size table

Studies that reported reduced well-being among preterm individuals were conducted with children who exhibited high neonatal clinical risk, such as GA < 32 weeks [14, 16–18, 31, 33], birth weight <1500 g [17, 33], or admission to the Neonatal Intensive Care Unit (NICU) [17, 31]. The presence of cognitive or functional disabilities [14, 16–18] and behavioral problems [16] at school age (6–11 years of age) in preterm children was also noted. The study by Ketharanathan et al. [32] was distinctive in being the only one to report no statistically significant difference between the parents’ reports of QoL of their children (aged 2–5 years) who were born preterm compared with age-matched controls who were born full term. This may be explained by the fact that the preterm children in Ketharanathan et al. [32] had a higher GA, had not required admission to the NICU, and did not show any major behavioral problems.

Schiariti et al. [31] compared parents’ reports of QoL in groups of preterm individuals who were stratified by GA. They found that QoL at 3 years of age in individuals with a GA of 28–32 weeks was similar to the group with a GA < 28 weeks. However, both groups of preterm individuals were at a disadvantage compared with the QoL reported by the parents of full-term children. Multivariate analyses suggested that better QoL outcomes were associated with the highest GA among groups of preterm individuals with a GA < 32 weeks [33] or 32–35 weeks [35]. Thus, the effect of the level of prematurity on QoL in preterm children requires further investigation.

Clinical risk factors associated with worse parent-reported QoL at preschool age in preterm children included non-lethal congenital malformations [33] and the use of continuous positive airway pressure ventilation support during the neonatal period [32]. Better parent-reported QoL at 5 years of age in children born preterm was associated with multiple pregnancies [33]. Furthermore, receiving immunoprophylaxis against respiratory syncytial virus infection (as per medical recommendation) before the age of 3 months was a protective factor for parent-reported QoL during the first year of life [35]. Results were inconclusive regarding the association between QoL and birth weight status or gender [14, 16, 17, 32, 33].

Better parent-reported QoL at 1 [35] or 6–10 years of age [14] was associated with several psychosocial factors including having siblings, parental mental health, and a reduced burden of care on parents. However, risk factors associated with the worst parental and self-reported QoL at school age (6–11 years of age) in preterm children included belonging to more disadvantaged social groups [14, 17, 18] and black race [17]. Findings were inconsistent regarding the parents’ schooling and employment status [16, 18, 33].

School-age children’s cognitive level was measured using the intelligence quotient (IQ) assessed with the Kaufman Assessment Battery for Children. Study authors considered different cutoff points of IQ < 85 [17], IQ < 81 [19], and IQ < 70 [16] to delimit low cognitive level. Nevertheless, low cognitive level was consistently associated with poorer parent-reported QoL [16, 19] and self-reported QoL [17]. Likewise, Berbis et al. [14] identified reduced parent-reported QoL in children at 6–10 years of age with global developmental disability, cerebral palsy, epilepsy, and hearing or visual impairments. Additionally, negative QoL outcomes were also reported by parents in school-age children with behavioral problems [16] or psychiatric disorders [19].

The only study that used both self- and parental reports of QoL evaluated children at 8 years of age and reported low levels of agreement between the two ratings in all QoL domains [17]. Parents reported worse QoL in various domains, whereas children reported the same QoL as their full-term peers. Using the Child Health and Illness Profile-Child Edition, preterm children reported less comfort and resilience than their parents but better performance at school and better peer relationships. These findings held true even when excluding premature individuals who had sensorineural disability [17].

Quality of life in adolescents born preterm

Table 3 shows the six articles on QoL in premature individuals in the adolescent age group (13–18 years of age). Five studies (83 %) in this age group compared self-reported QoL between individuals born preterm and a full-term control group [20, 22–25]. The vast majority of these studies reported that preterm adolescents reported satisfactory QoL, of a similar level to healthy full-term adolescents, both in high-risk samples [22, 24, 25] and those with low neonatal clinical risk [23].

Table 3 Characteristics of the studies and summary of main results related to QoL in adolescents born preterm (13–18 years of age; n = 6)
Full size table

However, Wolke et al. [20] reported that individuals born very preterm who did not have any sensorineural disabilities had worse QoL than full-term adolescents, regardless of the source of information (parents or adolescents themselves). Despite the overall agreement between parents and adolescents [20], consistency between ratings depended on the QoL domain being assessed using the Health Utilities Index Mark 3. Agreement was higher in domains related to physical health (vision, dexterity, ambulation, and hearing), but lower in domains related to psychological functioning (expressive language, emotion, cognition, and pain). Generally, adolescents reported more problems in language, cognition, and pain than their parents did. According to the authors, this negative view of adolescents about their own cognitive and emotional aspects may be related to this particular developmental phase, which includes more social and educational requirements [20].

Mid-late adolescence (14–19 years of age) did not affect self-reported QoL of preterm individuals [21]. Worse self-reported QoL in adolescents born preterm with a low birth weight was associated with physical disabilities at 5 years of age [21], as well as low IQ, language delays, emotional problems, cerebral palsy, and hearing or visual impairments at 8 years of age [20]. Similarly, worse QoL was related to factors in adolescence that may negatively affect perceptions of QoL during that same period, such as internalizing behavior disorders and neuromotor disabilities [21].

Quality of life in adults born preterm

Table 4 shows a summary of the six studies that assessed QoL in preterm individuals assessed as adults (19–44 years of age). All five studies comparing self-reported QoL in preterm adults with full-term adults reported favorable results with similar QoL [27–30, 34]. Furthermore, preterm adults reported better QoL in three domains of the SF-36, even after adjusting for gender and clinical variables [29]. Despite these positive outcomes, preterm individuals compared with full-term adults were still more likely to have chronic diseases [28], internalizing behavior disorders, fewer interactions with friends, and lower self-esteem related to sports and social acceptance [27].

Table 4 Characteristics of the studies and summary of main results related to QoL of adults born preterm (19–44 years of age; n = 6)
Full size table

The only domain of the SF-36 in which worse QoL was reported among adults who were born preterm was “Mental Health” [27]. However, three other studies using the same instrument did not find this difference [28, 29, 34]. Notably, the adults in the study by Lund et al. [27] generally had more health problems.

Disparate results were reported for possible factors, such as gender, that might be associated with QoL in preterm individuals as adults. Baumgardt et al. [28] reported worse self-reported QoL in men than in women, whereas Ulrich et al. [30] reported no significant interaction between gender and general well-being. Many other clinical and sociodemographic factors investigated were not related to QoL outcomes in adulthood [30, 34].

Discussion

In this review, we investigated associations between prematurity and QoL. Similarly to a previous systematic review by Zwicker and Harris [9], we found apparent improvements in QoL in preterm individuals as they get older, and this finding was unrelated to birth weight. Unlike the previous review, here we used more studies with samples of preterm individuals for whom a reference range of birth weight was not established. As well as reaffirming previous findings, we furthered knowledge by investigating possible explanations for this difference in QoL between age groups. We identified and analyzed risk and protective factors associated with QoL and considered the sources of information used by the studies.

Preterm children generally had lower QoL than children who were born full term. However, adolescents and adults who were born preterm had comparable QoL scores to full-term individuals. These results should be considered cautiously because of the great variety of instruments and methods used in the studies. Moreover, QoL is a subjective construct that can change according to a person’s stage of life as wishes, necessities, and demands change over time [8]. Additionally, individuals with a history of various risk factors may develop resilience to overcome such adversity. Many protective factors can coexist and consequently offset adversity faced in childhood, creating a positive adaptive outcome [38].

It is possible that differences in QoL between age groups may not only be due to developmental changes, but also the particular source of information. Parents and guardians were the main informants for children’s QoL in younger age groups, whereas studies with adolescents and adults were based on self-report. Parents of children with health problems tend to underestimate the QoL of their offspring [39]. Therefore, one should consider the influence of feelings, desires, and personal perceptions in analyses of parent-reported measures of a child’s QoL. For such parents and guardians, the birth of their children can be a particularly stressful time, which may influence their perception. Nonetheless, parental reports are still relevant because they have a major influence on children’s education and identification of possible health problems [40].

Agreement between parental and children reports of QoL in preterm individuals was lower in the younger age group compared with adolescents. Parents reported worse QoL in domains related to physical health, whereas preterm individuals reported worse QoL in mental and emotional health domains. This suggests that parents of individuals born premature emphasize aspects related to physical health when determining the QoL of their children. However, these assertions are based on the analysis of only two studies that evaluated agreement between parents and children. Therefore, this issue requires further exploration and confirmation in other samples of preterm individuals. Similar conclusions were drawn in a previous systematic review that sought to verify the relationship between parent- and self-reported QoL in children with chronic health problems [40]. Parents and children seem to agree on subjects related to physical health, which involve more observable aspects of a child’s life (illness, limitation of daily activities, etc.). However, there was high disagreement in QoL related to social and emotional functioning, which are more subjective domains, including personal feelings and perceptions.

These findings, along with our own, suggest that for samples of individuals born preterm, QoL assessments in children should consider both parent- and self-reported measures [40]. Discrepancies between the sources of information should be considered when discussing results, with a focus on disagreements between parents and children in specific QoL domains.

Several other factors besides premature birth were associated with QoL and may explain the differences between age groups. QoL was associated with clinical factors during the perinatal and neonatal phases and sociodemographic factors only in individuals who were born preterm and evaluated in childhood. However, some studies with adolescents or adults also included these as possible predictors of QoL and found no such association. This suggests that neonatal clinical risk conditions are associated with QoL specifically in early development. Moreover, these conditions could lead to parents having a worse perception of the QoL of their vulnerable children.

Problems that are often seen in individuals who are born very preterm (e.g., cognitive deficits, behavioral problems, and sensorineural or physical disabilities) were consistently associated with poorer QoL outcomes in children and adolescents who were born under high-risk conditions. However, most of these cognitive and behavioral disadvantages do not appear to be related to QoL in older age groups. It must therefore be considered how these individuals can report satisfactory QoL in adulthood despite the difficulties they faced.

This review also aimed to critically analyze the methodological quality of the included observational studies. The majority of studies showed good methodological quality, using standardized, accurate, and precise instruments, which reduces the risk of bias in our findings. Moreover, few studies were identified as having specific methodological problems. However, in one cross-sectional study, confounding variables were not controlled for in the analysis [35], and in two studies, instruments were used that had psychometric properties that had not yet been published [14, 30].

Previous studies of QoL of preterm individuals have focused on high-risk samples; therefore, future research should assess samples of preterm individuals with low clinical risk or who are defined as moderate or late preterm. Studies should also consider self-reported QoL of children and further investigate the influence of GA, birth weight, gender, parental schooling, and employment status on QoL of preterm individuals from different age groups. Factors influencing QoL of preterm individuals in adulthood should be better explored. These analyses should be combined with longitudinal assessments of QoL using repeated measures, which would permit observations of changes over time. We suggest that more complex statistical analyses should be conducted to identify predictive models that can evaluate mediating and moderating effects on QoL outcomes. A previous systematic review by Mottram and Holt [10] on the influence of GA on QoL recommended that future studies could also use qualitative designs. Considering the findings of the present systematic review, this recommendation has not yet been adopted by researchers. Finally, the present review highlights the relevance of including QoL assessments in studies of preventive and therapeutic interventions with individuals born preterm. We recommend that clinical trials control for factors associated with QoL outcomes indicated in observational studies, such as sociodemographic factors, cognitive level, behavior problems, and neurosensory impairments.

Conclusion

Discrepancies were found between parent- and self-reported QoL of individuals born preterm dependent on age group and source of information. In childhood, parents of preterm children reported worse QoL compared with parents of full-term children, whereas adolescents or adults who were born preterm self-reported more positive QoL outcomes. Several risk factors associated with preterm birth could explain such differences. Neonatal risk conditions and developmental problems during childhood may be related to parents’ negative QoL. Nevertheless, the factors that might be associated with QoL in individuals born preterm require further investigation, especially in adulthood.