Review questions

The research questions that this systematic review aims to address, are:

1. to assess the prevalence and the incidence in Italy of pediatric conditions approved as indication for GH treatment: GHD, TS, CKF, PWS, SGA, SHOX gene deficiency;
2. to assess adherence to GH treatment in Italy and identify factors that could facilitate or hamper the adherence;
3. to assess the economic impact of GH treatment in Italy;
4. to assess the quality of life of patients treated with GH and their caregivers in Italy.

Searches

Systematic searches have been performed in PubMed, Embase and Web of Science, looking for papers published from January 2010 to March 2021, describing national or international studies. Only articles written in English or Italian were included. The complete search strategies are reported in S1 Table.

Inclusion criteria

Types of study to be included

• Research question 1 (epidemiology): epidemiological studies assessing the prevalence/incidence of the conditions with an indication for the treatment with GH in pediatrics;
• Research question 2 (adherence): studies of any design assessing adherence to GH treatment;
• Research question 3 (economic impact): economic studies assessing the economic impact of the treatment with GH, in terms of total costs, cost-effectiveness, cost-consequence, cost-utility, drug wastage costs, and other relevant economic measures.
• Research question 4 (quality of life): studies of any design assessing the quality of life of patients treated with GH and their caregivers.

For all the research questions, only studies performed in Italy, or international multicenter studies with at least one Italian center, were included.

Main outcomes

• Research question 1 (epidemiology): prevalence and incidence of conditions treated with GH in Italy; in absence of Italian estimates, worldwide estimates were considered;
• Research question 2 (adherence): adherence to GH treatment;
• Research question 3 (economic impact): cost-effectiveness and other economic analyses of GH treatment;
• Research question 4 (quality of life): quality of life of patients treated with GH and their caregivers.

Measures of effect

• Research question 1 (epidemiology): prevalence and incidence;
• Research question 2 (adherence): medication possession ratio (MPR), proportion of days covered (PDC), proportion of doses administered vs prescribed, percentage of adherent patients;
• Research question 3 (economic impact): total costs, drug wastage costs, incremental cost-effectiveness ratio (ICER), quality-adjusted life year (QALY), etc.;
• Research question 4 (quality of life): results reported in qualitative studies (questionnaires, interviews, etc.) aimed to assess the quality of life in patients treated with GH and their caregivers.

Literature screening process

The study selection process was performed by two independent reviewers. Any disagreement was solved through discussion and, when necessary, a third reviewer was contacted. Study selection was conducted in two phases. Initially, the reviewers assessed the records through the titles and abstracts screening, according to the predefined inclusion criteria. In the second phase, the reviewers evaluated the full-text of the potential eligible studies. The final studies included in the review were described in the main text and in the tables, while a list of excluded studies along with the reasons for exclusion have been published as (S2 Table). Bibliographic references were managed using the EndNote X7.4 software.

Data extraction

Data extraction was performed by one reviewer and verified by another reviewer using a standardized form. The following information was extracted from the included studies: bibliographic data (first author, publication year and title), study characteristics (study design, location, study period, number of centers, sample size), participant characteristics (disease/condition, gender, age), treatment characteristics (type of GH, dosage and duration), comparator characteristics (type of comparator, dosage and duration), study outcomes (incidence/prevalence, adherence measures, economic measures, quality of life scores).

Specifically, for each research question we extracted the following information.

Research question 1 (epidemiology)—secondary studies: number of included studies and their references, disease, epidemiological estimates (incidence, prevalence, etc.); primary studies: disease, study population (N), epidemiological estimates. None of the primary studies identified in this review are in common with those included in the secondary studies.

Research question 2 (adherence)—definition of non-adherence / non-adherent patients (%), reasons for non-adherence, factors influencing (barriers / facilitators) the adherence.

Research question 3 (economic impact)—type of economic analysis, type of costs / discount rate / economic perspective / reference year, results.

Research question 4 (quality of life)—methods used for assessing QoL, subjects interviewed, results.

Quality assessment of included studies

The quality assessment was performed by one reviewer and verified by another reviewer, using appropriate checklists according to the study design of included studies. In particular, the following checklists were used: for prevalence studies, the checklist by Hoy et al. [1]; for cohort studies, the CASP checklist [2]; for qualitative studies, the CASP checklist [3]; for case series, the Institute of Health Economics (IHE) Quality Appraisal Checklist for Case Series Studies [4]; economic studies were evaluated by the Consensus Health Economic Criteria (CHEC) list [5].

Strategy for data synthesis

According to study protocol, a quantitative synthesis of the results through a meta-analysis would have been performed only in the case three or more (included) studies reported on the same outcome for any research question; moreover, to be combined, the studies would have to be similar in terms of PICO (patient/population, intervention, comparator, outcome). Assuming a high level of heterogeneity between the included studies, it was planned to perform random effects meta-analyses, providing cumulative estimates along with 95% confidence intervals. The heterogeneity between the included studies would have been assessed using the I² statistic, considering a statistical significance level of p <0.05 for all analyses. The STATA 13/SE software was indicated for the statistical analyses. It was also specified in the protocol that, in all the cases where a meta-analysis was not considered feasible, the results would have been presented narratively.

Epidemiology

To address the research question 1 (epidemiology), we identified 9 studies [7, 13, 14, 17, 19, 25, 28–30], 5 of which were secondary studies [14, 25, 28–30] and 4 were primary studies [7, 13, 17, 19] reporting epidemiological data about the considered diseases. The main characteristics of the included studies are described in the Table 1 (secondary studies) and Table 2 (primary studies).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of included secondary studies for epidemiology.

https://doi.org/10.1371/journal.pone.0264403.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of included primary studies for epidemiology.

https://doi.org/10.1371/journal.pone.0264403.t002

Among the secondary studies, we included 3 narrative reviews [14, 25, 29], 1 economic analysis with a narrative review on epidemiology [28], and 1 commentary [30].

The 2 reviews by Cicognani [14] and Nicolosi [25] reported the prevalence of SHOX deficiency in ISS children. The review by Cicognani [14] included 7 studies [31–37], 6 of which reported international estimates [31–34, 36, 37], and 1 reported Italian estimates [35]. Most of the included studies were in common between the two reviews. Cicognani [14] reported a prevalence range of 1.1%-14%, with an expected prevalence in children of 1/2,000, while Nicolosi [25] reported a prevalence range of 1.1%-15%, with an expected prevalence in children of 1/1,000.

Other two included studies were those of Tornese [29, 30], related to SGA. The first article [29] was a narrative review about SGA, in which it is stated that, by definition, SGA children should be 2% of the population (− 2 SDS of birth weight/length, corresponding to the 2nd percentile). This hypothetical prevalence is lower than the prevalence reported in the three population studies included in the review: 3.1% (Finland) [41] (SGA was defined as– 2 SDS of birth weight compared to a reference population), 5.5% (Sweden) [42], and 3.5% (Japan) [40] (in the last two studies, SGA was defined as − 2 SDS of birth weight or birth length compared to a reference population). In the population study conducted in Sweden, Karlberg et al. [42] reported that 12% of the children with SGA had growth retardation at 2 years of age. According to this ratio, considering a minimum prevalence of 2% of SGA in the general population, the hypothetical prevalence of short stature in SGA children would be 0.24% (1: 417) at two years of age. A Japanese study [40] verified the prevalence of short-stature children born SGA by studying a cohort of nearly 30,000 children born over a three-year period and re-evaluated at 3 years of age: the prevalence was 0.06% (1: 1,800).

The other study by Tornese et al. [30] was a commentary based on data from the Italian National Registry of Growth Hormone therapy (RNAOC), in the years 2011–2017. This study showed that the prevalence of SGA patients treated in 2017 was 0.37/100,000 (1.79/100,000 considering children 0–15 years). The incidence of treated SGA patients was 0.42/100,000 per year (mean of the years 2011–2017). The authors commented this data as extremely lower than expected, concluding that many short children born SGA are still not properly identified in Italy, and therefore not treated with GH, or misdiagnosed as GHD.

Another secondary study included was the review paper by Spandonaro et al. [28], aimed to determine how much of the variability in GH consumption in Italy is actually due to differences in clinical practice, and how much to waste. This literature review identified 6 articles [42–47] reporting epidemiological data on 5 GH-treated conditions; the epidemiological estimates are showed in Table 1.

In addition, our literature search retrieved 4 primary epidemiological studies [7, 13, 17, 19] (Table 2). Behnisch et al. [7] analyzed a cohort of 594 patients with CKD from 12 European countries; the prevalence of growth failure varied between countries from 7 to 44%, while the prevalence in Italy was 23.9%.

Chiavaroli et al. [13] described a retrospective review of obstetric records over a 20-year period in the Chieti province (Italy). The incidence of SGA children was stable in the study period, ranging from 8.3% in 1993 to 7.6% in 2013. Prevalence in this study is higher compared to studies described above, due to a different definition of SGA (birthweight and/or length <10th percentile, i.e. <−1.28 SD).

The paper by Genoni et al. [17] described a prospective cohort study conducted in Novara (Italy) in the period 2012–2015. This study found a prevalence of SHOX-D of 6.8% in a population of 281 children with short stature.

In another epidemiological study included, Kodra et al. [19] presented data on PWS from 2 Italian registry, the Italian National Rare Diseases Registry (NRDR) and the National Hospital Discharge Database (HDD). The annual incidence of PWS in a 3-year period (2012–2014) was 0.08 and 0.10 per 100,000 in data from NRDR and from HDD, respectively.

In summary, in the included studies:

• the estimated prevalence of GHD in the pediatric population (international estimates) is about 1/4,000–10,000, while the prevalence of GHD in Italy was indirectly estimated through the index of exposure to GH treatment in the Piedmont region (Italy), that is 9.44 subjects per 10,000 residents <18 years;
• the minimum hypothetical prevalence of SGA children is 2%, with a range of 3.1–5.5%; the hypothetical prevalence of SGA children with growth retardation at 2 years is 0.24%, while in a Japanese study it was 0.06%;
• the prevalence of SHOX-D in children with idiopathic short stature ranges between 1.1–15.0%, while prevalence in the general population of children is about 1/1,000-1/2,000 (international estimates);
• the birth prevalence of Turner syndrome is 1/2,500 (international estimates);
• the birth prevalence of PWS is about 1/15,000 (international estimates), while the annual incidence in Italy is reported between 0.08 and 0.10 per 100,000 in registry studies;
• the birth prevalence of CKD in the general population is 0.001%, and the prevalence of growth failure among CKD children in Italy is 23.9%.

Adherence

To address research question 2 (adherence), we identified 10 studies [6, 8–12, 18, 20, 22, 23], 5 of which report on adherence to GH treatment in general, using different approaches to assess adherence [6, 9, 10, 18, 23], and 5 studies describing the adherence with the Easypod™ device (Saizen®, Merck Serono S.A.—Geneva, Switzerland) [8, 11, 12, 20, 22], an electronic auto-injector that automatically records the patient’s adherence to treatment. The characteristics of included studies are reported in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of included studies for adherence.

https://doi.org/10.1371/journal.pone.0264403.t003

The multicenter study by Bagnasco [6] was conducted in 46 Italian pediatric centers, with the aim to evaluate self-reported adherence to GH therapy in a representative sample of Italian children and adolescents, and to identify the determinants of poor adherence. Non-adherent patients, defined as those that missed ≥1 injection over a typical week during the last 12 months of GH treatment, were 24.4%. The main reasons reported for non-adherence are: being away from home (33.3%), forgetfulness (24.7%), not feeling well (12.9%), and pain (10.3%). Among the barriers to adherence, that is, factors associated with low adherence, low level of parent education and longer treatment duration are reported. Among the facilitators, that is, factors associated with high adherence, convenience of the injection device and awareness of the consequences of non-adherence are reported.

The international, multicenter study by Bozzola et al. [8] was conducted in 206 centers in 15 countries, with the aim to assess the opinions of users of the Easypod^TM auto-injector and the adherence to the treatment. Non-adherent patients, defined as those that missed more than two daily injections per month or six daily injections in a 3-month period, were 12.5% of the entire study population (n = 824; GHD 66.4%, TS 9.8%, CRF 1.7%, SGA 15.3%, other 6.8%), of which 10.3% were naïve, and 18.3% experienced users. With regard to the Italian centers involved, non-adherent patients were 20.9% (15.4% naïve vs 34.4% experienced). Among the reasons for poor adherence, the following have been reported: forgetfulness (43.7%), device malfunction (18.2%), running out of cartridges/needles (12.9%), being away from home (12.6%).

Another study by Bozzola et al. [9] aimed to examine adherence to GH therapy in GHD patients in whom a growth deficit has been observed after a period of good response to therapy. In addition, new strategies are proposed for the management of non-adherence to long-term GH therapy. Among 11/106 children in whom growth failure was observed, 10 admitted poor adherence. Among patients, there were different levels of non-adherence, from missing occasional doses per week to discontinuing the therapy. Reported reasons for non-adherence included complex treatment regimen and pain.

Buzi et al. [10] carried out a narrative review of international studies investigating adherence to GH treatment. The authors reported that the definitions of non-adherence and the methods used for its assessment vary considerably between the included studies. Non-adherent patients ranged from 5% to 82% in the different countries. Among the barriers to adherence, the following were reported: discomfort, complex therapeutic regimens, age, personal factors, level of understanding the benefits of treatment. Among the facilitators: use of automatic injection devices or increasingly fine needles, use of needle-free devices.

The paper by Cardinale et al. [11] describes a multicenter study conducted in 6 Italian centers, with the aim of evaluating adherence to GH treatment in patients with growth deficiency, monitored using the Easypod ™ device. Adherence was defined as the ratio between actual days of treatment and planned days of treatment. Patients were classified according to different adherence levels: ≤50%, 50–60%, 60–70%, 70–80%, 80–90%. Non-adherent patients were on average 30% of the total.

Centonze et al. [12] conducted a multicenter study in 22 Italian centers. The study reported the 3-year prospective adherence data of the Italian cohort of naïve GHD children extrapolated from the Easypod Connect Observational Study (ECOS) database. Adherence is defined as the percentage of injections recorded by the device compared to those prescribed. The mean adherence was > 85% (1^st year: 88.5%; 2^nd year: 86.6%; 3^rd year 85.7%).

The paper by Giavoli [18] described the results of a survey carried out during the period of the COVID-19 emergency among patients treated with GH and their families. The objective of the survey was to collect information on the clinical conditions of the patients, on their laboratory results, as well as reinforce a recommendation to stay at home as indicated in the emergency phase. In addition, possible changes in adherence to therapies during the pandemic period were verified. In the pediatric population, adherence was measured by the Morisky Medication Adherence Scale (MMAS); in this scale, a total score of 8 indicates a high level of adherence, a score between 6.0 and 7.9 a medium level of adherence, and a score less than 6 a low level of adherence. The adherence of adults/transition age patients was self-reported: patients who declared taking > 80% of the total number of the prescribed GH injections were considered adherent.

In the pediatric GHD population, 82% of patients showed high adherence, while in the adult/transition population 94% declared an acceptable adherence. The reasons for poor adherence in the pediatric population were related to drug supply problems and motivational problems, while in the adult/transition population mainly drug supply problems.

The paper by Loche et al. [20] described a multicenter study conducted in 10 Italian centers, with the aim of evaluating adherence in GHD patients treated with GH using Easypod™. Adherence was defined as the proportion of injections correctly administered during the observational period out of the expected total number of injections; patients were considered fully adherent if they had an adherence rate ≥ 92%. Adherence data was available for 53/79 children: 30/53 reported ≥300 injections in the 12 months of the study; 17/30 (56.7%) had an adherence rate of ≥ 92%.

Maggio et al. [22] conducted a single-center, observational, retrospective study based on real-world data, with the aim of describing the correlation between the effectiveness of GH treatment administered via Easypod ™ and adherence. The average adherence (% of injections administered vs prescribed) in the children involved was 92.2%. Adherence was inversely correlated with the patient’s age, the duration of therapy, the number of GH doses. In the study, height gain did not reach a significant correlation with treatment adherence.

The paper by Marcianò [23] reported a drug-utilization study using administrative databases. The aim of the study was to explore the pattern of use of biosimilar and originator GH in six Italian centers. The study showed that 2,428 of 4,493 naïve patients (54.0%) discontinued therapy for at least 60 days over the 6-year study period. There were no statistically significant differences in persistence to treatment between biosimilar GH and originator. Most of those who discontinued were intermittent users (39.3%), that is, they restarted GH therapy after at least 60 days of interruption; in their paper, Marcianò et al. [23] have speculated that treatment discontinuation could be due to patients’ reduced compliance or lack of perceived benefit from GH therapy. Another possible reason for discontinuing GH therapy may be the occurrence of adverse reactions; in fact, more than 25% of patients who discontinued therapy had a new diagnosis of diabetes or cancer or has been hospitalized at least once after discontinuation. Another reason for discontinuation could be reaching final height (7.9% discontinued at age 15–17). The remainder of the subjects who discontinued therapy (27.7%) were thought to be probably related to the lack or loss of GH efficacy or to patients’ lack of compliance.

Economic impact

To address the research question 3 (economic impact), we identified 5 studies [15, 16, 21, 26, 28]. The characteristics of included studies are reported in the Table 4. One of these studies [28] has already been considered for the research question about epidemiology, while another study [21] has been also considered for the research question about quality of life.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Characteristics of included studies for economic impact.

https://doi.org/10.1371/journal.pone.0264403.t004

The paper by Drube et al. [15] described an international consensus statement on GH treatment in children with chronic kidney disease and includes a cost-effectiveness analysis based on the median price of one gram of GH in eight European countries, including Italy. This study showed the total drug-related costs for four different scenarios: 1) 5-year-old patient at the start of treatment and 2-year treatment duration: total cost € 13,000; 2) 5-year-old patient at the start of treatment and 5-year treatment duration: total cost € 37,900; 3) 12-year-old patient at the start of treatment and 2 years of treatment: total cost € 27,100; 4) 12-year-old patient at the start of treatment and duration of treatment of 5 years: total cost € 80,100. The corresponding incremental cost per centimeter gained in height as an adult for the 4 scenarios was, respectively, € 1,800, € 5,300, € 3,800 and € 11,100.

In the study by Foo et al. [16], patients who received somatropin with Easypod ™ gained, on average, 3.2 cm more than patients who received other GH treatments. In the base case analysis, which considered the list prices of drugs, somatropin administered with the Easypod ™ device (Saizen®) was the second in terms of average total cost for a complete multi-year GH treatment, including the cost deriving from drug wastage (Omnitrope® € 57,024, NutropinAq ® € 73,304, Humatrope® € 78,098, Norditropin SimpleXx® € 90,689, Genotropin® € 96,188, Saizen® € 96,710, Zomacton® € 104,895), but it was the second least expensive per cm gained (Omnitrope® € 6,058/cm, Saizen® € 7,699/cm vs NutropinAq® € 7,787/cm, Humatrope® € 8,297/cm, Norditropin SimpleXx® € 9,634/cm, Genotropin® € 10,218/cm, Zomacton® € 11,143/cm). Considering published tender prices (scenario analysis), Saizen® had the lowest cost per cm gained (Saizen® € 4,708/cm, Omnitrope® € 5,065/cm, NutropinAq® € 6,179/cm, Humatrope® € 6,442/cm, Genotropin® € 6,528/cm, Norditropin SimpleXx® € 6,917/cm, Zomacton® € 9,851/cm) across all compared treatments.

The paper by Lopez-Bastida et al. [21] aimed to determine the economic burden from a societal perspective and the health-related quality of life (HRQOL) of patients with PWS in Europe. A survey was conducted on 261 patients (48 in Italy) to estimate the social/economic costs including three dimensions of costs: direct healthcare costs, direct non-healthcare costs (formal and informal care) and labour productivity losses. The study results per patients < 18 years (175 in total, 32 in Italy) showed that the average annual costs ranged from € 3,937 to € 70,083 between countries (€ 33,787 in Italy). Direct healthcare costs ranged from € 458 to € 16,695 (€ 4,974 in Italy), and direct non-healthcare costs ranged from € 1,387 to € 52,389 (€ 28,813 in Italy).

Pagani et al. conducted a study [26] including two groups of subjects: 12 children without GH deficiency (bioinactive GH) and 12 children with GH deficiency; both groups were exposed to GH treatment. The study found a significant (p <0.05) increase in growth rate in both groups during the first year of GH treatment (non-GHD: –1.7 to 5.4 standard deviation score (SDS); GHD: –1.5 to 4.7 SDS). There was no statistically significant variation between the two groups in the difference between final height and target height. Similarly, no significant difference in cost/height gain was found between GHD children (€ 1,925 ± 653) and children with bioinactive GH (€ 1,640 ± 631). Furthermore, no significant differences were found in the annual cost of therapy between GHD (€ 12,348 ± 2,018) and children with bioinactive GH (€ 11,355 ± 1,748).

The study by Spandonaro et al. [28] was conducted to determine how much variability in GH consumption in Italy is actually due to differences in clinical practice and how much waste is related to the devices used for drug administration. The model estimated that, for some devices, waste can amount to up to 15% of consumption. The consumption model, applied to the Italian population, has shown that in various Italian regions one can assume either an excessive prescription or a potential under-prescription, ranging from 20 to 40% less to over 200% more than the estimated theoretical consumption.

Quality of life

To address the research question 4 (quality of life), we identified 4 studies [8, 21, 24, 27]. Another study [48] describing the validation of the Quality of Life in Short Stature Youth (QoLISSY) questionnaire in Italy was selected but not included in the systematic review since it does not report quality of life results; this study is briefly described in the Discussion section. The characteristics of included studies are reported in Table 5. One of these studies [8] has already been considered for the research question about adherence, while another study [21] has been also considered for the research question about economic impact.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Characteristics of included studies for quality of life.

https://doi.org/10.1371/journal.pone.0264403.t005

The two studies by Lopez-Bastida [21] and Ragusa [27] focus mainly on the disease (PWS) rather than on the GH treatment. We decided to include them in our analysis because we were interested in analyzing the quality of life in patients with PWS, which is one of the diseases for which GH treatment is approved in Italy, and we assumed that most patients were treated. In the Lopez-Bastida paper [21], information on medications used by PWS patients was obtained from questionnaires and their costs calculated. In the Ragusa study [27] there were references to GH treatment in interviews with various patients and caregivers.

In the international multicenter study of Bozzola et al. (2011) [8] which included 206 centers in 15 countries, including Italy, the Easypod ™ device was found to be acceptable by users. The device was described by 89.1% (716/804) of the participants as convenient/very comfortable to use, and 91.8% (732/797) of respondents said they would like to continue using the device.

The paper by Lopez-Bastida [21], previously described in this review, showed that in adult PWS patients the mean EuroQol 5-domain (EQ-5D) index score (where 0 corresponds to death and 1 corresponds to perfect health), ranged between 0.40 and 0.81 (Italy: 0.40), while the mean EQ-5D visual analogue scale score (where 0 corresponds to worst imaginable health state and 100 to best imaginable health state) ranged between 51.25 and 62.63 (Italy: 56.15). Although this study included patients with a mean age of 14 years (range 6–22), quality of life measures have been reported only for adult patients. Among caregivers, the mean EQ-5D index score ranged from 0.73 to 0.82 (Italy: 0.82), and the mean EQ-5D visual analogue scale ranged from 70.26 to 81.52 (Italy: 77.81).

The qualitative study by Marini et al. [24], conducted in 11 Italian centers on patients with GHD and their caregivers highlighted a general satisfaction among patients with regard to social and school life and GH treatment outcomes, while there was a certain level of intolerance to GH treatment among adolescents. One-third of the adolescents said about the GH treatment “I don’t like it but I have to do it”; 39% of them lived the medical visits with boredom and impatience, and 50% of them suffered for the daily task. Most of parents (67%) had concern of not being well informed about side effects. Both the GHD patients and their parents appreciated the work of healthcare professionals and were satisfied with the outcomes of therapy. The narratives of healthcare professionals showed the need to find a more engaging way of communication with adolescents, and to reassure the families about therapy and its possible effects, in order to increase the adherence to the therapy.

The multicenter study by Ragusa [27] et al., which involved 10 Italian centers, investigated the impact of PWS on patients and caregivers through narrative medicine, to understand the problems related to daily life, the needs and resources of PWS patients and their caregivers and provide insights into clinical practice. Disbelief, displacement, anger and pain represented the most recurrent emotions expressed by caregivers in an attempt to adapt to the situation and its criticalities. From the caregivers’ narratives, food-seeking behaviors emerged as the most challenging event to manage in the home context. Relations outside the family were difficult to preserve, imposing a radical change in social life. Sixty-two percent of family caregivers had to change job after the birth of a child with PWS.

Strengths and limitations

The main strengths of this review are: it is based on a study protocol registered in PROSPERO; the reporting of the review follows the PRISMA guidelines; a comprehensive literature search was performed in three electronic databases; the literature selection process, data extraction and quality assessment were performed independently by two reviewers.

Potential limitations regard the time frame considered for the literature search (2010–2021) and that we considered only published studies, without searching the grey literature; these limits could have reduced the sensitivity of the search.