The diagram in the infographic above comes from:
- Estes, 2016: Maternal immune activation: implications for neuropsychiatric disorders https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5650490/
The diagram in the infographic above comes from:
The studies cited by the CDC on their “Vaccine Do Not Cause Autism” page cannot possibly support that claim. The CDC’s conclusion is invalid.
See the infographic above for details about why that is.
Our understanding and conception of autism has changed considerably over the years, as shown in the infographic above.
The first recognition of autism only occurred in 1943, and for decades it was believed to be very rare, affecting only 1-in-10,000 children. The first time the American Psychiatric Association (APA) published their list of all known mental disorders, the prestigious Diagnostic and Statistical Manual of Mental Disorders or DSM-1 (1952), it did not include autism. It used the word autistic only in reference to one of the behaviours associated with childhood-onset schizophrenia. Autism was also not listed in the second edition, DSM-2 (1968).
Autism was first listed as a known mental disorder in DSM-3 in 1980. A category of neurodevelopmental (early onset) disorders was created called the Pervasive Development Disorders (PDDs), which contained infantile autism for when autistic behaviours appeared before 30 months of age and childhood onset PDD for when they appear after 30 months of age.
These autistic behaviors include lack of responsiveness to other people, deficits in language development and social relationships, resistance to change, attachment to inanimate objects, anxiety and panic attacks, inappropriate fear or rage reactions, oddities of motor movement, and speech and hyper-sensitivity to sensory stimulation.
A third PDD was atypical autism, a catch-all for when there were impairments in social skills and language but the criteria for a specific PDD cannot be met.
The DSM-4 in 1994 introduced three areas of diagnosis for the PDDs: impairments in social interaction, impairments in communication, and restricted, repetitive and stereotyped patterns of behavior, interests and activities. It also renamed and reformulated the PDDs.
Autistic Disorder required onset before 36 months of age and at least 6 of the 12 impairments listed across the three areas of diagnosis. Those previously diagnosed with infantile autism would in the new system likely be diagnosed with autistic disorder. Those previously diagnosed with childhood onset PDD or atypical autism would likely be diagnosed with PDD-NOS (for Not Otherwise Specified), the new catch-all for where there are severe deficits in one of the three areas of diagnosis but the criteria are not met for any of the specific PDDs.
Three new PDDs were added: Asperger’s Disorder was created for those with impairments in social interaction and restricted behaviours, but no impairments in communication, cognition or language skills, and it was considered “mild autism” or “high-functioning autism”. The other two new additions were Childhood Disintegrative Disorder (CDD), a sudden and severe regressive disorder in multiple areas of functioning with subsequent onset of autistic behaviors, and Rett’s Disorder, a regressive disorder affecting head growth, motor skills, gait and trunk movements, as well as autistic behaviours. These are both very rare disorders, and Rett’s Disorder is now considered a genetic brain disorder rather than an autism spectrum disorder or PDD.
Asperger’s makes up 17% of cases of PDD, autistic disorder is 31%, but the largest group is PDD-NOS at 53% (Hviid, 2019). The addition of Asperger’s Disorder (and to a much lesser extent, CDD and Rett’s) would thus have caused only a small step change – ceteris paribus – in the number of people diagnosed with a PDD, as DSM-4 replaced DSM-3.
The DSM-5 in 2013 replaced the PDD category entirely and replaced it with Autism Spectrum Disorder (ASD). The PDDs were already considered a spectrum widely known as the autism spectrum, ranging from the most severe (Autistic Disorder) to the less severe (PDD-NOS and Asperger’s). Autism was no longer the name of one type of PDD, but the whole category, and it was defined as a single spectrum disorder rather than a category of related disorders with distinct names, as PDD had been.
Within ASD, there is a classification by severity in two dimensions: 1) deficits in social interaction and communication, and 2) restricted, repetitive behavior, interests, or activities. Three severity levels (mild, moderate and severe) are defined by how much support the individual requires. Most of those previously diagnosed with autistic disorder would now be diagnosed with moderate or severe ASD (aka autism), while most of those previously diagnosed with Asperger’s disorder would now be diagnosed with mild or moderate autism. It is possible to be diagnosed as having one severity level for social deficits and a different severity level for restricted behaviors, giving nine possible severity level combinations within the spectrum.
The DSM-5 also created a new neurodevelopmental disorder called Social (Pragmatic) Communication Disorder (SCD) for cases that meet the social deficits criteria for ASD but not the restricted behaviors criteria. According to one study (Kim, 2014), only 63% of those with a previous diagnosis of PDD-NOS meet the DSM-5 criteria for ASD, with 32% lacking the restrictive behaviours criteria and so would now be diagnosed with SCD, not ASD. Those previously diagnosed with PDD-NOS who do meet the ASD criteria are likely to have either a mild or moderate DSM-5 severity level (Walker, 2004).
The impact of a third of those with PDD-NOS (the largest PDD disorder) being taken out of the umbrella of ASD should have a noticeable decrease – ceteris paribus – in autism diagnoses as DSM-5 replaces DSM-4. The definition of ASD in DSM-5 is tighter than the definition of PDD in DSM-4 and may be tighter than the definition of PDD in DSM-3 as well.
The World Health Organisation (WHO) maintains a comprehensive medical coding system known as the International Classification of Diseases (ICD). The current iteration of this system is ICD-10, which has been used since 1994. Chapter 5 contains the codes for mental and behavioural disorders (F00-F99) and it has codes for “childhood autism” (F84.0), “atypical autism” (F84.1) and Asperger’s disorder (F84.5), “Other PDD” (F84.8) and “Unspecified PDD” (F84.9) all within the category of pervasive development disorders (F84).
These five ICD-10 codes are considered to cover the whole autism spectrum and epidemiological studies of autism generally define an outcome of autism as a diagnosis with one of these five ICD-10 codes (for example, Hviid 2019).
These codes map easily to the disorders in DSM-4 and DSM-3: childhood autism is the same as infantile autism (DSM-3) and autistic disorder (DSM-4); Asperger’s Disorder is defined the same way as in DSM-4; PDD-NOS (DSM-4) is split between ICD codes for atypical autism, other PDD and unspecified PDD, with atypical autism being defined in a way similar to the DSM-3 definition.
Despite the many taxonomical changes within the PDD / ASD category of disorders over the years, the diagnostic criteria have been refined and elaborated with each iteration of the DSM, but the basic idea has always been the same. Autism has always been defined as a disorder with two fundamental defining elements: deficits in social communication and social interaction, and restricted, repetitive patterns of behavior, interests and activities.
DSM-5 gives the clearest diagnostic criteria for each of these two elements. It provides three illustrative examples of social deficits, all of which must be met for a diagnosis, and four illustrative examples of restricted behaviors, two of which must be met for a diagnosis. The examples are:
Persistent deficits in social communication and social interaction across multiple contexts, as manifested by ALL OF the following, currently or by history:
Restricted, repetitive patterns of behavior, interests, or activities as manifested by AT LEAST TWO of the following, currently or by history:
For a diagnosis of ASD, symptoms must also be present in the early developmental period (but may not become fully manifest until social demands exceed limited capacities, or may be masked by learned strategies in later life), they must cause clinically significant impairment in social, occupational, or other important areas of current functioning, and these disturbances are not better explained by intellectual disability or global developmental delay.
The requirement for onset during the “early developmental period,” meaning early childhood, obviously means that adults cannot develop autism. However, adults can be and are diagnosed with autism if they are impaired significantly enough by their social deficits and repetitive behaviors today to warrant a diagnosis, without any evidence that their autistic behavior started in the early developmental period. The age of onset diagnosis criteria is essentially ignored as it is assumed that any adult that meets the criteria for autism today must have been autistic their whole life.
The requirement for “clinically significant impairment” is standard wording throughout the DSM-5 for all mental disorders. It means that it is left entirely at the discretion of clinicians to judge whether a child (or adult) is impaired *significantly enough* by his social deficits and repetitive behaviors to get a diagnosis of ASD. Clinicians get to draw the line between marginal cases of mild ASD and no ASD, and where they draw that line is flexible and may have changed over time to include more people: an adult previously considered merely as having autistic-like traits (perhaps an introvert with restricted interests, a geek or an eccentric) may now get a diagnosis of ASD, if he seeks one. There are incentives for adults to get a diagnosis, for parents and schools to have their children diagnosed, and for clinicians to give diagnoses in marginal cases. There is a clear danger of over-diagnosis of autism due to the flexibility of the clinical significance requirement, and this must be borne in mind when looking at historical statistics of autism prevalence and incidence.
Clear and well-understood definitions are essential for productive discussion and good science. When discussing autism, it is important that all the participants understand whether the term is being used:
The wider sense includes those with Asperger’s (17%) and PDD-NOS (53%), which are likely mild or moderate autism in the DSM-5 conception.
When discussing autism, when reviewing data from epidemiological studies, comorbidity studies, historical trends, independence surveys, and so on, it is crucial to understand if the data refers to the whole autism spectrum or just those with severe autism, or those with autism and some other condition comorbid with autism, such as a language disorder or intellectual disability.
The Institute of Medicine (IOM) systematic review entitled “The Childhood Immunization Schedule and Safety: Stakeholder Concerns, Scientific Evidence, and Future Studies (2013)” confirmed there had been no studies of the vaccine schedule, and it called for such studies to be done.
It then tells us the most feasible way to carry out these urgently-needed studies:
The most feasible approach to studying the safety of the childhood immunization schedule is through analyses of data obtained by VSD. VSD is a collaborative effort between CDC and 9 managed care organizations that maintain a large database of linked data for monitoring immunization safety and studying potential rare and serious adverse events. VSD member sites include data for more than 9 million children and adults receiving vaccinations on a variety of immunization schedules.
The VSD (Vaccine Safety Datalink) is potentially a goldmine of data that could be decisive in the vaccine science debates, but the CDC keeps it locked up. It makes the data available only to select individuals; it is not publicly available for independent researchers to analyse.
Making anonymised VSD data available to everyone would be an easy and cheap way to enable epidemiological studies of all different vaccine schedules to be carried by anyone who has doubts about vaccine safety or efficacy and wants to verify the raw data.
Why doesn’t the CDC want independent researchers or parents to be able to compare health outcomes between populations vaccinated on different schedules or unvaccinated?
The vaccine schedule has not been tested.
This is according to the Institute of Medicine (IOM) systematic review entitled “The Childhood Immunization Schedule and Safety: Stakeholder Concerns, Scientific Evidence, and Future Studies (2013)”, a report which is cited by the CDC.
Here is the full quote:
In summary, few studies have comprehensively assessed the association between the entire immunization schedule or variations in the overall schedule and categories of health outcomes, and no study has directly examined health outcomes and stakeholder concerns in precisely the way that the committee was charged to address in its statement of task. No studies have compared the differences in health outcomes that some stakeholders questioned between entirely unimmunized populations of children and fully immunized children. Experts who addressed the committee pointed not to a body of evidence that had been overlooked but rather to the fact that existing research has not been designed to test the entire immunization schedule.
Vaccines have been tested, but the vaccine schedule has not been.
A test of the vaccine schedule would entail comparing health outcomes between populations given different combinations of vaccines, including fully unvaccinated, selectively vaccinated, and fully vaccinated according to the recommended schedule (which is different in different countries). As the IOM quote above makes clear, there are no such studies.
Retrospective cohort and case-control studies comparing health outcomes between populations vaccinated using different schedules are urgently needed.
In 1915, in England & Wales, there were 16,455 deaths from the measles; 70% of measles deaths were children aged 1-4. Measles was the biggest killer of toddlers, responsible for 20% of all toddler deaths. It was also the 2nd biggest killer (behind diphtheria) of children aged 5-9. Although 1915 was a peak year, the average at the time was around 10k total measles deaths per year.
In 1994, there were no measles deaths at all. From then until now, there have been few years where we have had more than one measles death.
How did we achieve this? How did we conquer measles?
The answer can be found in this chart that shows the number of measles deaths (black) between 1901 and 2016, and the number of measles notifications (red) between 1940 and 2016.
From 1915 to 1955, measles deaths reduced by 99%, from over 10,000 deaths per year to below 100 deaths per year.
Most of the decline happened between 1915 and 1935. In 1935, there were only 1,346 deaths from measles, a reduction of 92% over a 20-year period. The measles death rate fell even faster than the overall infant mortality rate; it was now responsible for only 7% of deaths of children aged 1-4 (now only the 3rd biggest killer), and only 2% of deaths of children aged 5-9 (now only the 10th biggest killer).
In 1955, there were only 176 deaths from measles, a reduction of 99% over a 40-year period. In most years of the 1960s, there were fewer than 100 measles deaths per year, and measles was no longer among the top 10 killers of children of any age group. Measles was now responsible for only 3% of deaths of children aged 1-4 (the 7th biggest killer) and 2% of deaths of children aged 5-9 (the 9th biggest killer).
Turning now to measles notifications, for which we have data only back to 1940, we can see that the number of notifications of measles was flat from 1940 to 1968 (the pre-vaccine period), at around 400k on average. The driver behind the dramatic decrease in the number of deaths was therefore a dramatic reduction to the Case Fatality Rate (CFR), i.e. the number of deaths per case (or the inverse of the survival rate). Just as many children caught measles, but far fewer died from it.
We can use notifications as a proxy for cases to calculate the CFR for the period after 1940. To estimate the CFR for the period 1901-1940 we need to use birth numbers, as in the following chart:
By assuming a similar proportion of children caught the measles before 1940 as in 1940-68 (i.e. most of them), we can extrapolate the CFR backwards, giving us the following chart:
The CFR prior to 1915 was over 2%, meaning that 1-in-50 children who caught measles died from it. By 1940, it had decreased to 0.2% (1-in-500), and by 1955 it had decreased to just 0.02% (1-in-5000). This is a 99% reduction in the CFR. This reduction in the CFR is why measles was no longer considered a serious, life-threatening illness during the period 1955 to 1968; 99.98% of children survived it. The average number of measles deaths per year was down to below 100 by 1955.
In 1955 the incredible reduction in the CFR suddenly stopped. It remains at around 0.02% even today, as can be seen in this chart:
Let us now re-scale the deaths axis of our first chart to see what this meant for the deaths figures after 1955:
As a result of the flat CFR, the number of deaths stayed at just below 100 per year for the period 1955-1968. The record lowest numbers of measles deaths in the pre-vaccine era were in 1956 (30 deaths), 1960 (31 deaths), 1962 (39 deaths). The most deaths in this period were in 1961 (152 deaths), 1963 (127 deaths), and 1965 (115 deaths) and no other years in this period had more than 100 deaths.
From 1968 to 1997 there was a 99% reduction in measles cases. The majority of that reduction followed the introductions of the measles vaccine (1968) and the MMR vaccine (1988). From 1968 to 1971, the annual number of measles notifications fell by over 60%, from 400k to 150k. From 1988 to 1991, notifications fell by 90%, from 100k to 10k. With a flat CFR, the 99% reduction in cases from 1968 to 1997 meant a 99% reduction in deaths, from 100 per year before vaccines down to just 1 per year.
In most years since 1997, there have been under 5k cases and rarely more than 1 death, as we can see by one final re-scaling of the chart:
Measles killed over 10,000 per year before 1915 but fewer than 100 per year by 1955. This 99% reduction in measles deaths was due to a dramatic reduction in the deadliness of the disease between 1915 and 1955. This could have been due to a healthier environment (higher standards of hygiene, better sanitation systems, more nutritious food, cleaner drinking water and air, less cramped living and working conditions, etc) or to better treatment of measles cases (improved medical knowledge among doctors and the public, better access to healthcare, more effective quarantine procedures, etc), or both.
The healthier environment and better quality treatments deserve credit for making measles a relatively mild disease; these developments saved thousands of lives each year, just in England & Wales. The measles and MMR vaccines reduced the number of measles cases and measles deaths by another 99%, saving dozens more lives per year. Vaccines deserve credit for making it rare for anyone to suffer from the measles.