Nosokinetics

Nosokinetics (Service care delivering modelling) is a term used to describe the science/subject of measuring and modelling the process of care in health and social care systems.

Contents 1 Modeling 1.1 Why model 1.2 Modelling limitations 1.3 Modelling strengths 1.4 Model types 1.4.1 Stocastic models 2 External links

Modeling

Why model

• Prediction
• Objectivity to decisions
• Consequences of decisions
• Identifying what is important
• Performance & monitoring measures
• Explaining what went wrong

Modelling limitations

• Measurable parameters must exist
• Assumptions
  • Usually include political and social stability
• Chaos limitations
  • War, earthquake or that unexpected epidemic

Modelling strengths

• Can often use data that is routinely collected
• Assumption examination can allow what if scenario planning
• Can allow for regular or repeating in time events such as seasonal influenza, effect of weekends and public holidays
• Can allow for demography changes

Model types

• Stochastic
  • Assume you do not know precisely what will happen in advance, but past experience and reasonable population size allow assumption of randomness that underlies many real-world phenomena
  • Allows a range of estimates
  • More useful when you want a background level of guaranteed service
  • More complicated so less likely to appeal to decision makers
• Deterministic
  • Returns a most likely estimate
  • Is easier to grasp but is far more likely to lead to chaos if used in prediction
    • Health and social care decisions based on deterministic modelling have created major system crises in most such systems
      • Examples include:
        • Creating assessment panels or other rationing (queuing) steps such as waiting lists for inpatient investigations in patients who are already inpatients.
        • Decreasing social service funding while increasing health funding leading to game playing by the underfunded service

Stocastic models

• Will tend to have to use phase-type (PH) distributions such as length of stay in an institution
  • State transitions need the mathematical concept of the (finite-state) continuous-time Markov chain (CTMC) which is rendered vis vector arithmetic.
  • Coxian distributions, a subtype of phase-type distributions are useful in healthcare modelling
    • Care home residents classically have a two state coxian distribution to their length of stay, because a subgroup rapidly die (or move on to more dependent institutional care) as they have a subacute unstable medical condition while the other subgroup have a chronic and only slowly progressive condition.
  • General phase-type (PH) distributions would be better for modelling patient flow around a hospital since the processes in the emergency department, theatres and wards are so different with the potential for readmissions and indeed use of complex numbers to =best fit the observed distrution of length of stay.

External links

• Nosokinetics Group web site
• NHS institute modelling and simulation tools