Chronic Pain
-pain persisting beyond 3 months, this is beyond the expected time of tissue healing
-the nervous system has changed, pain persists even when the original injury is gone, nonfucntional pain
Acute Pain
a signal, protective, useful, expected to resolve as tissue heals
Pain as Perception
-shaped by the brain, not just by tissue, this is why psychological and social factors matter so much
-without the brain you have no painful experience, based on physiological but also personality, mood, SES, gender and many other factors interacting in the brain
Chronic Pain Prevelance
-1 in 5 adults worldwide live with chronic pain
-its the leading cause of disability
-more years lived with disability than any other condition globally
Economic Burden
-estimated $635 billion annually in the US alone, more than cancer, heart disease, and diabetes combined
Opioid Use & Chronic Pain
-chronic pain is the driver of opioid prescribing and a key factor in the opioid crisis
Yakoi Study Problem
-most psych studies test whether associations exist
-they dont test whether models can forecast new cases
-the replication crisis is largely a prediction failure
Yakoi Study Core Idea
-prediction and explanation are different goals requiring different methods
-a model that explains may be useless at predicting
-out-of-sample accuracy is the true test of a model
-more data beats better theory when the goal is prediction
Yakoi Study Solution
-use cross-validation, test your model on data it has never seen
-use large datasets
-change the question: can I predict pain for this person
Prediction Goals
-prediction refers to the capacity of a model to predict disease states from high dimensional data
-the problem is that many studies have claims to establish prediction while only providing correlation
-the goal is to generalize these predictions is new individuals that were never encountered before
Steps toward Prediction Goals
-predict the risk of developing pain or worsening of their pain
-improve personalized treatment
-improve allocation of health care resources
-refine diagnoses and phenotypes
-provide mechanistic insights
-pain clinics have wait times of 1-2 yrs
-these methods can expand diagnoses and can improve mechanistic insights to find targets for interventions
-there is a mosaic of factors that shapes pain in each individual
Risk Factor
something measurable before chronic pain develops that increases its likelihood
Differences in Pain
-same injury->persistent, disabling pain
-leads to depression, lost work, opioid use for high-impact pain
-same injury->recovery within weeks
-returns to work, normal life
-different types of pain, even if at the same pain level, have different meanings (3/10 stomach pain is not experienced the same as 3/10 cancer pain)
Predicts Worse Outcomes
-fear avoidance behaviours
-psychiatric comorbidities
-high baseline pain
-low general health status
Predicts Recovery
-low fear avoidance
-good baseline function
-strong social support
-positive recovery expectations
Biopsychosocial Model
biology, psychology, social context
Biology
genetics, sex, neurological sensitivity, inflammatory markers
Sex Differences
women report chronic pain more often and show greater pain sensitivity on lab tests, not fully explained by hormones alone
Genetic Variance
-genes influence pain sensitivity, polygenic factors, opioid response and susceptibility to chronic pain
-each variant has a tiny effect, thousands of ppl are needed to detect them reliably
Brain Imaging
-brain based biomarkers predicting subjective experience of pain has become an obsession for the field, with limited results
-not the most ethical to research, but would be helpful to locate pain objectively
-current resting scanning does not have enough insight
Psychology
fear avoidance, catasrophisizing, depression, anxiety
Fear Avoidance
avoiding movement that could cause harm or worsen the pain
-lose physical fitness and functional decline over time
-high fear avoidance roughly double the odds of a poor outcome at one year
Social Context
work environment, SES, healthcare access
Reductionisim Problem
-building simple models that appear theoretically elegant but have limited capacity to predict actual human behaviour
-developing complex models that can accurately predict bahviour but fail to respect known psychological or neurobiological constraints
-using machine learning to find intersection between findings in blood, genes, bone, and brain to go beyond the constraints to optimize predictions
