This is a question that no Nurse or Regulatory Inspector has ever been able to answer in spite of the fact that the use of probability for single event management is a complete nonsense.
The responses to date to this question are best described as varying degrees of hostility, declarations that it is integral to evidenced based care practice and emphatic disinterest in addressing this gross misuse of probability.
This silliness would be laughable if the reality of this practice was not so serious for patient care and the vast waste of organisational resources that it is responsible for.
Worryingly, this misuse of probability is being enshrined in Regulatory Standards e.g. the Health Information Quality Authority in Ireland requires care providers to use assessment formats so that the probability of each resident / patient in regard to falls can be predicted and measures put in place to prevent and or reduce the chance of a fall happening.
I am not an expert, my competency in statistics is limited to three years of Economic and Social statistics as minors within my BSc Joint Honours majors of Sociology and Social & Economic History but that is sufficient to understand these basic principles:
PROBABILITY: the following are gross misuses:
a) Trying to predict 'when' - especially low-occurrence high impact events
b) Using the past to manage / predict a future
This misuse of probability is exacerbated by a demonstrable lack of understanding of the difference between an INDICATOR and a MEASURE.
In 2013 N.I.C.E. issued guidance that nurses should not use assessment tools that purport to measure (e.g. High, Medium, Low Risk) patient's probability of experiencing an event (in this case falls); guidance that is equally valid for hospitals and care homes.
Yet the daily norm for thousands of nurses in Public (e.g. NHS) organisations and Private care homes is spending huge amounts of time at a desk completing predictive assessments rather than in hands-on patient care.
prof. dr. Patty Brennan has done research is using prediction models back in the late 1980ies. I took a course in 1989 in which she addressed using prediction models in nursing. It is a way back, but most fundamental research of high quality has been undertaken decades ago.
Further to go into depth into the polemic, it would help to get examples how the use of probalistic analysis should and should not be used. The situation above describes the use cases. However, how would that apply in the real world of the use of the probability calculation? And why should we not use probability for predictions? Seems that not only nursing is doing this.
Thank you William for taking the time to submit a view.
My question was raised in this forum to facilitate an open discourse to facilitate a better understanding of the limitations of probability so that initiatives to achieve better safety for patients is realigned on developing real-time interventions to reduce vulnerability of individuals to the impact (harm in this case) of relevant risks rather than spending vast amounts of valuable direct care opportunities sitting behind a desk filling out predictive falls assessment tools that both common sense can see and the plethora of published research has found to be of no use.
In regard to your last comment questioning why we should not use probability, my point is not against appropriate uses of probability but a specific misuse i.e. that of nurses trying to predict when a single fall event will happen and preventing its occurrence.
I believe this resource may provide some insight into the direction of patient probability models in the STEADI algorithm, which is incorporated into many patient electronic health records systems.
http://gerontologist.oxfordjournals.org/content/early/2016/04/29/geront.gnw074.full
Thank you Sanden for taking the time to submit a contribution.
Perhaps clarification is needed about this issue that concerns me i.e. the use by nurses of what are called predictive assessments in which the goal is to PREVENT a fall falling by working though a format that ascribes a risk / probability value that is typically identified as a patient as being at low, medium or high risk of falling.
There are many straightforward hazard elimination and passive impact provision measures that will mitigate harm to patients from the falls; an outcome that cannot be achieved by the use of predictive assessments.
These references, one of many credible sources, provide good background about the issue concerning predictive falls assessment tools:
- UK's National Institute for Health & Care Excellence (N.I.C.E.) guidance reference 1.2.1.1 in June 2013: Do not use fall risk prediction tools to predict inpatients' risk of falling in hospital (This guidance is equally applicable to long term care facilities like nursing homes).
- This link http://ageing.oxfordjournals.org/content/37/3/248.long will take you to Professor Oliver and his team's research which provides extensive clarification about the transparent lack of veracity of all predictive fall assessment tools and the misleading outcomes that they can cause.
It is not possible to use probability for single event management i.e. in this case preventing a fall by predicting who will fall and when. However and in spite of the obvious uselessness of predictive falls assessment tools, nurses across the UK and Ireland spend vast amounts of their time off-the-floor away from patients completing predict assessments.
The obvious risk statement and associated real-time helpful interventions are ignored:
- Everyone (staff and patients) in a hospital can fall for all sorts of reasons
- Therefore, fall safety interventions should involve a combination of
- Trip and Fall Hazard elimination where possible
- Person Specific Hazard Identification and Risk Response Strategy
- Situation Specific location / environmental hazard assessment and risk management strategy e.g. there may well be a number of differing risk response protocols in Intensive Care to those in an Ambulatory Day Ward in regard to a Fire.
- Badges
- Science topic
- Similar topics
- Medicine
- Public Health