Predictors of Outcome in Autism Spectrum Disorders: A Perspective for Clinicians and Therapists

Professionals working with families who have individuals with Autism Spectrum Disorder (ASD) are frequently asked about the possible long-term outcome of their children and the duration of intervention needed to ‘move out of the spectrum’.^1,2 Clinicians and therapists often have to predict the subsequent outcome of an individual child based on all the services offered.^1,3

Also, families who get a diagnosis of ASD in the Indian context for their child typically raise the question of whether the child will do well and recover completely. They wish to know if the child can attend mainstream school, be self-sufficient, be employed, or get married like most others in the community. This is a natural process of exploration in the pathways to care that most parents go through until they have met many professionals, understood ASD, discussed it with other parents, and started working with their child pragmatically. Predicting outcomes with the data available or acquired during assessments and intervention predominantly focuses on the short term. This can be best defined as improvement in specified skill sets across one-to-two-year durations. However, parents naturally expect an understanding of long-term outcomes across the lifespan.

In this viewpoint, we attempt to present our viewpoint based on our experiences working in Indian academic institutions and hospitals and from the literature review to summarize possible predictors of outcomes that can be useful for clinicians and therapists.

We recognize the view of the neurodiversity advocates disappointed by the focus on therapies that work towards normative outcomes for persons with an ASD diagnosis. We recognize the spirit of such advocacy. However, framing outcome issues from that perspective is beyond the scope of this viewpoint. Here, we focus on the frequently noted requests from the families.

What Is ‘Outcome’ in the ASD Context?

In the context of ASD, it is important to understand what ‘outcome’ means. Is outcome (1) A set of skills the child acquires as a natural phenomenon when left without intervention (naturalistic outcome)? (2) A phenomenon of development of various skills, which results as an output of intervention by multiple stakeholders such as therapist, individual with autism, parents, and school (interventional outcome)? (3) A prediction of what will happen shortly based on the current-day assessment and intervention with the child (predictive outcome)?^4–6 Ideally, much more research is needed to understand the outcome of the third type. Long-term outcome refers to outcomes beyond 5–10 years, usually focused on achieving equivalence with normative outcomes for the community.

What Factors Act as Predictors?

Predictors are baseline features that can reliably tell us how the future will be. When the clinician assesses the child, findings such as having good eye contact, smiling once during observation, doing a high five, pointing to the fan, or following commands help us understand how the child will respond to intervention. Then, they can help predict the clinical improvement in a year.^4,7

Therefore, predictors can help in two ways: (1) to predict the child’s clinical outcome by measuring their skills and (2) to predict if a particular intervention results in a particular outcome in the long term, that is, the functional outcome connected to intervention. Predictors vary with age. The importance of the presence of a particular skill at a particular age helps predict an outcome. We discuss the predictors as follows.

Child Factors

Predictive factors in children are predominantly behavioral observations in the child. These include speech, hyperactivity, repetitive behaviors, and being uncooperative. Apart from the observed factors, we also infer many aspects of a child’s behavior, such as being ‘bright/friendly’, adhering to a task, and following a routine. There can also be measured factors such as the scores on severity scales,^4,8,9 sleep index, ¹⁰ abilities across domains, ¹¹ developmental or intelligence quotient, ¹² adaptive behaviours, ¹² and comorbidities. ¹³ The observed, inferred, and measured factors can all be used to predict a given child’s outcome.^2,8,14–17

A comprehensive assessment at baseline helps make meaningful connections for future assessment models. Systematically measuring predictors is a starting point in predicting outcomes. We can assess severity (e.g., Indian Scale for Assessment of Autism—ISAA), development (e.g., Communication DEALL Developmental Checklist—CDDC/Intellectual Quotient—IQ), sensory processing (e.g., Sensory Profile—SP), language (e.g., Assessment of Language Development—ALD/Receptive-Expressive Emergent Language Test—REELS), carers report vs. direct assessments by the clinician, assessment of comorbidities such as sleep issues, epilepsy, ADHD, tics, and self-injurious behavior—SIBs.^18–21

It is also important to understand what the goal of the assessment is—prediction, certification, or planning intervention.

Parental Factors

This includes the aspects such as the ability to co-operate, adherence to therapy and schedules, ²² home-based intervention,^23,24 stress levels of parents,^24,25 marital discord, dysfunctional family, single parent, ²⁴ time taken to react to the diagnosis, and involvement with the therapist^26,27. A systematic review describes the impact of cultural beliefs on parenting stress, especially where ASD or related developmental disorders are considered a punishment to the family, resulting in stigma, shame, marginalization, and discrimination. ²⁸ A study from South Asia highlighted parental factors such as stress and caregiver burden in families because of stigma and lack of resources. ²⁹

Behavioral Predictors

Age- and context-dependent factors such as good eye contact, level of hyperactivity, imitation, and self-play can act as behavioral predictors.^30–32 However, in developing countries, data on measuring these specific features is usually lacking due to a shortage of professionals.

Language- and Communication-Based Predictors

Presence of verbal skills by four years indicate better language and communication skills by 10 or 15 years of age.^33,34 A child with narrative abilities and back-and-forth communication will have a better outcome than a child with responsive speech (responding only to questions asked). Comprehensive spoken and written language skills indicate a better prognosis than in those who have poor comprehension. A child who is also very good with nonverbal communication skills has a better prognosis than someone who is verbal but cannot communicate.

A meta-analysis found that on average, a third of children above five seemed to pick up language. ³⁵ It also found that the age at which you start intervention, baseline language ability, cognitive skills of the child, and length of follow-up did not influence the child’s outcomes. However, many studies that measured outcomes had various biases. Also, the above findings are contrary to the wider literature and the clinical experience at our centre.^34,36 Studies have found that baseline social and communication skills help predict one-year outcomes after intervention, irrespective of all other contributing confounders.^37,38 Cognitive skills at baseline also predicted later language production but not necessarily understanding language. Where the specialists such as speech therapists and developmental psychologists are unavailable, early intervention focused on reducing core autism symptoms is also very important to help develop later speech and language skills.^39,40

Developmental Predictors

Broad cognition of the child like IQ, ¹² social cognitive skills like social responsiveness scale (SRS), ⁴¹ and adaptive behaviors like those measured using Vineland adaptive behavior scale (VABS) or Vineland social maturity scale (VSMS).

Generally, better skills can point to a positive outcome. In preschool children with ASD, normal or near-normal performance can also predict a positive ASD outcome. The developmental trends can inform intervention planning. However, to predict long-term outcomes, it is important to consider verbal and non-verbal IQs, other observational measures of severity, and other variables.

Biological Factors

Though EEG-,^42–44 MRI-,^9,45–50 and genotype-based studies^51,52 have been conducted in the last decade, current international guidelines do not recommend using such parameters as part of routine clinical assessments in all cases. Currently, EEG, MRI, and genotype studies help to predict autism risk, symptom severity, language development and differentiate them from ADHD and other developmental disorders.^53–56 However, they do not yet have individual predictive value for each child. By and large, biological factors contribute to understanding short-term outcomes based on group data. More research is needed to determine if these have a translational benefit in predicting long-term outcomes.

Pharmacotherapy can be beneficial and could impact the course of ASD, especially in those with comorbid ADHD, epilepsy, self-injurious behaviors, aggression, anxiety, depression, sleep disorders, and so on. However, it is important to choose the right medication at a low dose, regularly monitor for outcomes, and withdraw as required. ⁵⁷

While all these factors help predict outcomes, substantial evidence indicates that early intervention positively impacts the developmental trajectory. Studies from across the globe have investigated the efficacy of these interventions.^40,41,58,59 It is well known that brain growth is rapid in the early years of life, with a window period during which intervention is expected to be significantly effective. ⁶⁰ 12–24 months of age is a critical period for brain plasticity and development. Intervention during this period facilitates better long-term outcomes.^61,62 If a child with ASD is brought later than 4 or 5 years of age, adequate evidence shows that mainstreaming the child is compromised to varying extents.^63,64 What was accepted to be a good time to start early intervention (3–4 years of age) two decades ago has now been proven otherwise by many longitudinal, family-based, high-risk ASD (siblings of children with ASD) studies. These studies opine that early intervention has to begin as early as 18–24 months to be effective and improve long-term outcomes.

It is important to define adequate early intervention by answering questions such as how early the intervention must be started; the nature, duration, and quality of intervention, and who is delivering the intervention (parents/professionals/multidisciplinary team). Duration of follow-up is also undeniably important and an important factor influencing long-term outcomes. Clinical observation during multiple time points, follow-up assessments, and understanding the neurobehavioral path an individual child with ASD has taken during a few years will help better understand and predict long-term outcomes.

Discussion

A variety of factors influence outcome. When predicting outcomes, it is important to record baseline scores of cognitive skills, adaptive behaviors, speech skills, and severity, and look at the rate of change in the scores over time during the intervention. Child factors, including social, adaptive, and language skills, absence of comorbidities, and overlapping comorbidities can predict a positive outcome. The true test of intervention is adequate follow-up, parental engagement, and preventing dropouts, which is the case in many research studies. It is appropriate to look at the hierarchy of outcomes such as self-care, adaptive behavior, especially social skills, language skills, academic targets, and independent living. It is essential to plan assessments and investigations to predict these outcomes that help in choosing interventions and to communicate progress to the families. It is also necessary to consider how frequent and well structured the assessments must be and can be.

Some of the challenges related to ASD are the progressive unfurling of autistic impairments, including but not limited to repetitive behaviors, as well as comorbid ADHD, anxiety, psychotic illness, and tic disorders, which are not related to early intervention. It is necessary to make this information understandable for parents and prepare them to face these challenges. We, therefore, must be conscious of the heterogeneities of the condition when predicting outcomes using our assessments and interventions. Choosing an intervention is always challenging as we do not have all facilities at all centers. Whether we decide to use an intervention based on literature, experience, or assessments, the capability to make a statement of positive predictability is challenging. Therefore, clinicians and therapists often need to decide upon outcomes that need more attention and focus on the interventions that are known to facilitate achieving those outcomes. The key, however, is to be mindful of some features which probably will not respond well to a particular kind of intervention offered by a therapist or a team but may benefit by probably going to someone else who delivers a different type of intervention. Currently, the findings are equivocal.^59,65

When we see short-term improvements, it may be clinical wisdom to expect and predict good long-term outcomes. If skills develop in the short term, it will set a cascade of reactions to help the child develop other skills not directly related to that intervention. For example, improved speech increases the likelihood of increased social opportunities, leading to potentially higher social skills.

Some good practices we should follow include (1) periodic assessments—global or specific; (2) working with a multidisciplinary team of specialists who can help with varied problem domains; (3) involving parents and keeping them motivated—parent/family assessments, providing family interventions, medical management of parents’ condition wherever indicated, and keeping parents engaged, informed, and prepared, and (4) providing intensive interventions.

Clinicians and therapists can first define what short-term outcomes must be for every child, evaluate outcomes of the intervention, and then look at long-term, life span course, and keep tweaking interventions along the way. It is essential to be systematic and conscious with interventions and when informing families with these predicted outcomes. Predictions, even if negative, should intensify work with the families.