Introduction

Polyhandicap is currently defined as a complex disability condition corresponding to a chronic affliction occurring in an immature brain, leading to the combination of severe/profound mental retardation and serious motor deficit and resulting in an extreme restriction with respect to autonomy and communication. Polyhandicap is similar to profound intellectual and multiple disabilities (PIMD). When the disorder affects an immature brain, the term of polyhandicap is used in France. Polyhandicap, as a subgroup of PIMD, includes the most severe cases due to the precocity of the brain affection1,2. In France, polyhandicap was initially provided by the French Polyhandicap Group (https://gpf.asso.fr/le-gpf/#definition); subsequently, it was retained in French law (Law No. 89–798 of October 27, 1989; Law No. 2005–102 of February 11, 2005; Decree of May, 9, 2017), and is now consensual within the scientific community3,4,5. To address methodological problems6, recently, the Ithaca European Reference Network (ERN) for congenital malformations and rare intellectual disabilities adopted the consensual term of PIMD/Polyhandicap2. In European countries, the prevalence of PIMD/Polyhandicap is estimated to be approximately 1 per 1000 7,8,9.

In France, individuals with PIMD/Polyhandicap receive care managed according to their health status severity10. The individuals presenting with the most severe health statuses and those requiring prolonged hospitalization are cared for in specialized rehabilitation centres that offer a high level of medical and paramedical physical rehabilitation and a high level of preventive care for medium-to-long-term stays. Individuals with polyhandicap who have less severe pathologies are cared for in residential facilities, which offer a high level of psychosocial education and a lower level of medical care. Some individuals, mostly children, receive care managed at home. Due to the consequences of neurodevelopmental retardation and the occurrence of comorbidities and/or impairments, individuals with polyhandicap may present with various disease courses, which can substantially differ from one patient to another11,12. Identifying predictive factors of health status change over time may be of interest for individuals with polyhandicap, their familles, institutional caregivers, and health policy-makers. It would help to better project in the future, to implement targeted (preventive and curative) care, and to enhance the care management environments.

Previous cross-sectional studies have described the health severity patterns of individuals with polyhandicap11,13. Longitudinal approaches can be used to provide a better understanding of the evolution of the polyhandicap course, thus enabling the exploration of dynamic sequences between them. The nationwide French EVAL-PLH (EVALuation PoLyHandicap) cohort study was designed to identify the socioeconomic, environmental, and epidemiological determinants of the health status of individuals with polyhandicap5. This is the first large cohort study focusing on polyhandicap using a consensus definition. No similar study has been conducted in France or in other countries. To date, two waves, spaced 5 years apart, were conducted. Many international publications reported the key findings from the first evaluation (2015–2016), highlighting the variety of health statuses according to age in terms of etiologies, comorbidities, associated handicaps, behavioural disorders, treatments and devices. The present longitudinal study (assessment 1: 2015–2016 and assessment 2: 2020–2021), examined a large sample of individuals with polyhandicap to: (i) describe the evolution over time of health status in terms of severity, (ii) and identify the potential predictors of health status change.

Materials and methods

General organization of the EVAL-PLH cohort

The French national EVAL-PLH cohort (clinical trial registration number NCT02400528, 27/03/2015), a prospective cohort study, was implemented in 2015. Data were collected at 2 points: (T1) the first wave was in 2015 and 2016, and (T2) the second wave was in 2020 and 2021. Individuals with severe polyhandicap were included. The individuals were cared for in various French centres such as specialized rehabilitation centres (4), residential facilities (9), and one specialized paediatric/neurological department corresponding to outpatients, i.e. people who receive care at home. Three different populations were eligible: (i) individuals with polyhandicap; (ii) familial referents of the included individuals (French legal mention for this type of person, represented by parents in the majority of cases); and (iii) institutional health care workers that treated the included individuals. The present study focused on the individuals with polyhandicap. The research was performed in accordance with the Declaration of Helsinki.

Selection criteria

The inclusion criteria for individuals with polyhandicap were as follows: age > 3 years (the diagnosis of polyhandicap and the neurodevelopmental assessment are accurate and reliable) at the time of inclusion (i.e. 2015–2016); presenting with a combination of motor deficiency (tetraparesia, hemiparesis, paraplegia, extrapyramidal syndrome, cerebellar syndrome, and/or neuromuscular problems), severe to profound intellectual impairment (intelligence quotient IQ < 40) associated with everyday life dependence (Functional Independency Measure (FIM) < 55), and restricted mobility Gross Motor Function Classification System (GMFCS) score of III, IV, and V); and age at onset of cerebral lesion was younger than 3 years old.

Data collection

Data were collected from medical records by a dedicated clinical research assistant, supervised by the referent physician of the patient (a referent physician is designated for each patient). At each wave, the following data were collected:

  • Sociodemographic data: age, sex, and structure of care management (specialized rehabilitation centre, residential facility, and home).

  • Aetiology: unknown or known; antenatal, perinatal, or post-natal: progressive or non-progressive.

  • Mobility status was assessed with GMFCS, autonomy status with FIM, and intellectual disability (ID) (profound ID defined by a neurodevelopmental age of less than 1 year vs. severe ID).

  • Neurodevelopmental patterns: The general neurodevelopmental status was assessed using the Brunet-Lézine scale, a developmental quotient assessment tool that can be used among infants up to 24 months old. While this tool cannot be used in patients over 2 years old, we scored 4 developmental domains (language, posture-motor abilities, coordination, and sociability), assuming that this scoring system provides an accurate estimation of the neurodevelopmental status. All scores ranged from 0 to 24 months14.

  • Behavioural disorders (clinical evaluation based on the presence of at least one of the following items: intermittent screaming and/or agitation and/or stereotypies and/or intermittent crying and/or self-aggression, and/or hetero-aggression).

  • Medical devices: number of devices (including invasive mechanical ventilation, noninvasive mechanical ventilation, tracheostomy, nasogastric tube, gastrostomy, permanent urinary probe, cerebrospinal fluid derivation, and central venous catheter).

  • Medications: number of medications (including laxatives, anti-convulsants agents, antalgics agents, psychotropics agents, antispastics agents, antidystonics agents, antibiotics, osteoporosis prevention, iron supplementation).

Definitions of the change in the health status over time

All the definitions were derived from the Polyhandicap Severity Scale (PSS), a widely validated 27-item tool15 assessing the severity of the health status. The scale provides 3 scores: an abilities score (according to the International Classification of Functioning, Disability and Health) ranging from 0 (best) to 60 (worse possible ability); a comorbidities/impairments score ranging from 0 (none) to 69 (more comorbidities/impairments), and a global score ranging from 0 (less) to 129 (more severe). Based on the 3 scores, each person was categorized into 2 classes according to the change in health status between T1 (inclusion) and T2 (5-year follow-up): non-worsened or stable-worsened.

  1. a.

    For the main analysis (primary definition): (i) non-worsened group: a person was defined as non-worsened when the score difference (T2–T1) was negative; (ii) stable-worsened group: a person was defined as stable-worsened when the score difference (T2–T1) was positive (> = 0).

  2. b.

    A sensitivity analysis was performed on a second definition: (i) non-worsened group: a person was defined as having a non-worsened health status when the score difference (T2–T1) was negative; (ii) stable-worsened-deceased group: a person was defined having a worsened health status when the score difference (T2–T1) was positive or a deceased person.

Ethics and regulations

Regulatory monitoring was performed in accordance with French law, which requires the approval of the French ethics committee (Comité de Protection des Personnes Sud Méditerranée V, Nice, France, 20/10/2014, Reference Number 2014-A00953-44). In accordance with French law, a written informed consent form was obtained from parents or guardians of the persons (minors and adults) included in the study.

Statistics

Descriptive data were provided as numbers and percentages for binary variables, and as medians and interquartile ranges for quantitative variables. For each PSS score, the proportions of individuals with a non-worsened or a worsened-stable evolution were determined based on the previously stated definitions. Bivariate associations between variables of interest and health status change (binary outcomes) were assessed using chi2 or Fisher exact tests, and the Mann-Whitney tests. Multivariate logistic regressions were performed to examine binary outcomes (the results are presented as odds ratios and 95% confidence intervals). The variables entered in the models were selected from the bivariate analyses (p < 0.05). Data were analysed by using SPSS 20.0 software.

Results

Characteristics of the sample

Among the 875 individuals included in the 2015–2016 wave, 155 individuals died between T1 and T2, 210 individuals were care managed in a nonparticipant centre at T2, 18 individuals were lost to follow-up, and 492 (56.2% of the initial sample) individuals underwent two assessments. The details are provided in the Fig. 1. The 492 persons who underwent two assessments were aged from 3 to 66 years at inclusion. Sixty-nine children (14.0%), i.e., individuals aged under 18 years at inclusion, became adults by the time of the 2nd assessment. The aetiology was unknown for 20% of the individuals. When the aetiology was known, the lesion occurred in antenatal time for 50% of the cases, perinatal for 26%, postnatal for 15%, and nonclassifiable for 7%. The aetiology was progressive for 18.8% of the sample. At T1, 54% of the individuals were cared for in specialized rehabilitation centres, 43% were cared for in residential facilities, and 4% were cared for at home. All the characteristics of the sample and the deceased persons are provided in Table 1. Compared to the 492 studied persons: (i) the 155 deceased persons did not differ on sex, age, and care management modalities; (ii) the 383 persons who did not undergo two assessments (including the 155 deceased persons) were significantly younger (years, 28 ± 16 and 20 ± 16, respectively, p < 0.001) and were less likely to receive care in specialized rehabilitation centres (54% vs. 38%, p < 0.001).

Fig. 1
figure 1

Chart.

Full size image
Table 1 Sample characteristics.
Full size table

Health status change from baseline to the 5-year follow-up

Table 1 shows the median scores and the interquartile ranges for the PSS scores at T1, and the PSS score differences between T1 and T2. According to the main definition, 68,8% (n = 309), 64.7% (n = 310), and 80.2% (n = 368) of the individuals presented with a worsened or stable severity status based on the global score, the abilities score, and the comorbidities score, respectively. Based on the definition for the sensitivity analysis, 76.8%, 73.3%, and 85.2% of persons were categorized as stable-worsened based on the global score, the abilities score, and the comorbidities score, respectively.

Predictive factors of a stable-worsened health status

For the main definition of the health status change (see above), the bivariate associations between baseline (T1: 2015-16) characteristics and health status change at 5 years (T2: 2020-21) are presented in Table 2. Multivariate analyses (Table 3) showed that: (i) the global score: the stable-worsened individuals, compared to the individuals with a non-worsened health status, were more likely to receive care in specialized rehabilitation centres, more likely to have an antenatal or progressive aetiology, and more likely to have a higher dependency level; (ii) the abilities score: the stable-worsened individuals, compared to the individuals with a non-worsened health status, were more likely to receive care in specialized rehabilitation centres, more likely to have an antenatal or progressive aetiology, more likely to have a profound intellectual impairment, and more likely to have a higher number of medications; (iii) the comorbidities score: no variable was linked to a stable-worsened health status.

Table 2 Associations between baseline (T1) characteristics and health status change at 5 years (T2): univariate analysis.
Full size table
Table 3 Associations between baseline characteristics (T1) and worsened health at 5 years (T2): multivariate analysis.
Full size table

When the deceased cases were included in the analysis as worsened cases (sensitivity analysis), the stable-worsened individuals were more likely to be cared for in specialized rehabilitation centres than the individuals with a non-worsened health status regarding the global score. An antenatal or progressive aetiology was associated with the worsened status. The results of the sensitivity analysis are presented in Table 4 (univariate analysis) and Table 5 (multivariate analysis).

Table 4 Associations between baseline (T1) characteristics and health status at 5 years (T2): the deceased persons were considered as part of the worsened cases: univariate analysis.
Full size table
Table 5 Associations between baseline (T1) characteristics and worsened health status at 5 years (T2): the deceased persons were considered as part of the worsened cases: multivariate analysis.
Full size table

Discussion

The present study was the first to provide a longitudinal description of the disease course of a large sample of persons with polyhandicap, as defined in France. To our knowledge, all previous studies focusing on the health characteristics of individuals with polyhandicap used a cross-sectional design11,12,13 and thus did not display the dynamic perspective.

The first important finding of this work is that 18% of the cohort died during the 5-year study period. While it is well-known that the life expectancy of people with polyhandicap has increased, especially due to the improvement of preventive treatments and enhanced knowledge of the specific needs of the individuals, it is essential that families understand that individuals with polyhandicap still die early11,16. In this specific population, there are no robust data regarding the estimation of mortality, median survival, and potential determinants of mortality. Future studies should explore these fundamental issues to better improve the care management of the patients.

However, even though the mortality rate remains substantial, this study yielded an encouraging finding. According to our definition, approximately one third of the sample of this study presented a non-worsened health status during the 5-year study period. This result contrasts with the common feeling of a perpetually and systematic worsening of the disease course of individuals with polyhandicap. Therefore, we can hypothesize that some individuals (here, 35%) experienced slight gains in some abilities due to rehabilitation and educational care, or improved their comorbidities score, including amenable disabilities, due to preventive and curative actions. Indeed, it is not uncommon to observe improvements in orthopaedic, digestive, dental, or other disabilities after surgery or rehabilitation or new treatments17,18,19. Here, again, it is necessary to spread this poorly known information among the health care community, and second among the families. Such information could change outlooks on the future, rekindle hope, and enhance caring.

The third important aspect of this work was the examination of potential predictive factors for health status deterioration over the 5-year period. First, whatever the definition (of worsened health) used, we found that the individuals cared for in settings dedicated to people with severe cases, i.e., the specialized rehabilitation centres, had the greatest level of health deterioration. In France, compared with residential facilities, it is generally agreed that the specialized rehabilitation centres offer a higher level of medical care (allowing, in particular, to increase life expectancy) and a lower level of educational care (mainly due to a lack of human resources). This may partly explain the association between a deterioration in health (especially the abilities score) and being cared for in specialized rehabilitation centers. A previous French study showed that the adequacy of care management was satisfactory from a health severity perspective10. Even if some discrepancies remain, the majority of the individuals with a severe health status were care managed for in appropriate settings, and the majority of individuals presenting with a less severe condition received care in outpatient settings. It remains challenging to improve the adequacy of care management from a clinical perspective, a setting (setting dedicated for children or adult people) perspective, or from the perspective of the families. This finding can also be used to discuss the issue of (de)institutionalization. Some benefits of the deinstitutionalization of persons with disabilities are now well established including improvements related to adaptive behavior, competence growth, community participation, and social interactions for patients and optimizing contact with family members20,21. Because in France deinstitutionalization of persons with PIMD is still a matter debate, this finding can help to consider the possible adaptation of this model to persons with polyhandicap22. Second, as expected, we found that individuals with an antenatal and-or progressive aetiology showed less favourable progress. Indeed, it is well known that a lesion occurring on the brain of a foetus always has a worse prognosis than a lesion occurring during or after the birth23,24. And, not in all models, a higher dependency and a profound intellectual impairment were linked to a worsened health over time. Third, less intuitively, the absence of behavioural disorders was paradoxically more common in individuals with a worsened severity status. One hypothesis is that behavioural disorders, such as auto-hetero-aggression or agitation, require motor skills and could thus not be present in the individuals with the most severe polyhandicap11. Fourth, the number of medications, as an indirect factor of severity, was occasionally associated with aggravation. This polypharmacy could not be thought of similarly to polymedication for older people. In older patients with polypharmacy, deprescribing is now recommended25 to reduce the likelihood of drug-drug interactions that can increase morbidity and mortality. In our specific context, the count was performed on treatments that cannot be designated as ‘comfort’ treatments (laxatives for example). These treatments are mostly not prescribed upon unreasonable request of the patient (or, more accurately, the caregiver), but the wellbeing of the patient is considered for some medications (anti-convulsants, antalgics, psychotropics, antispastics, antidystonics, antibiotics). Similarly, preventive medications (osteoporosis prevention, iron supplementation) also play an essential role in the reduction of comorbidities over time. Therefore, it can be difficult to follow the deprescribing routine. Nonetheless, awareness should be increased among caregivers. Finally, future studies should more carefully examine the role of amenable determinants that may assist the practitioners in their clinical practice.

This longitudinal study provides important information regarding health status changes among individuals with polyhandicap over a 5-year period. This knowledge is essential for families and health care providers to enhance their communication and improve their planning for the future. Even if an antenatal aetiology or a progressive disease were already perceived as pejorative elements, it can be more comfortable to rely on these objective results. Similarly, we hope that health decision-makers will use these findings. All things being equal, specialised rehabilitation centres seem be related to a worse evolution; therefore, it would be interesting rethink care management strategies for persons with polyhandicap as a whole.

Additional information will contribute to a better understanding of the course disease and the evolution of the entire target population: (i) The third wave of the EVAL-PLH cohort (planned for 2025–2026) will provide a more accurate overview of the determinants of health status changes; (ii) The sample did not include the deceased persons. This attrition bias could miss some determinants of health deterioration. Future studies should explore the profiles of deceased people; (iii) Other definitions of health status change should be explored; our definition was based on the Polyhandicap Severity Scale15, which is novel tool, that has not been used in previous studies, however, this scale is the only well-designed and well validated scale for measuring severity among individuals with polyhandicap persons; (iv) Future studies should explore medico-administrative databases. The French technical Agency ATIH edited recommendations for the identification of polyhandicap from the ICD (https://www.departement-information-medicale.com/wp-content/uploads/2013/08/Codage_Polyhandicap.pdf); it is essential to study trajectories of care management in an exhaustive sample of persons with polyhandicap; (v) The representativeness of the sample should be discussed. While approximately 70% of hospital beds dedicated to polyhandicap in France were located in the four specialized rehabilitation centres included in this study26, the participation of residential facilities, based on a voluntary basis, was underrepresented.

Conclusion

The current study provides a robust perspective of the worsened health evolution of persons with polyhandicap over a 5-year period. The most important predictive factors were the aetiology of the lesion and the care management modality. This knowledge could be important for families, health care providers, and health decision-makers.