Week 2: Appraising Evidence about Prognostic Factors

Prognosis

The process of predicting the future about a patient's or client's condition.

Secondary prevention

Decreases duration of illness, severity of disease, and number of sequelae through early diagnosis and prompt intervention.

Tertiary prevention

Limits the degree of disability and promotes rehabilitation and restoration of function in patients with chronic and irreversible diseases.

Primary prevention

Prevents a target condition in a susceptible or potentially susceptible population through such specific measures as general health promotion efforts.

Prognostic factor

A sociodemographic, diagnostic, or comorbid characteristic of an individual that confers increased or decreased chances of positive or adverse outcomes from a disease/disorder or from interventions.

The general term that describes 4 characteristics predictive of any type of future outcomes is

prognostic factors

Risk factors

Predictors of future adverse events usually are referred to as a sociodemographic, diagnostic, or comorbid characteristic of a patient or client that confers increased or decreased chances of development of an adverse outcome.

Survival curve

A graphic representation of the frequency of an outcome of interest over time created by plotting the percentage of individuals who are free of the outcome at successive points in time.

Odds ratio (OR)

The odds that an individual with a prognostic factor had an outcome of interest as compared to the odds for an individual without the prognostic factor.

Risk ratios

The ratio of the risk of developing an outcome in patients with a prognostic factor.

A ratio of the risk of the outcome in the experimental group relative to the risk of the outcome in the control group

risk factor

Hazard ratio (HR)

An estimate of the relative risk of developing the problem of interest over the course of the study, weighted by the number of subjects available.

P value

Indicates the probability that the result obtained occurred due to chance.

Confidence interval

A range of scores within which the true score for a variable is estimated to lie within a specified probability.

Case-control design

A retrospective research design used to evaluate the relationship between a potential exposure and an outcome; two groups of subjects - one of which has the outcome (the case) and the one that does not (the control) are compared to determine which group has a greater proportion of individuals with the exposure.

Inception cohort

A group of subjects that are followed over time starting early in the course of their disease or disorder.

Diagnosis

A process that labels patients, classifies illness, determines prognosis, and determines intervention.

Prognosis

A process that predicts...

Which outcomes could happen

The likelihood that outcomes will happen

The timeframe for outcome development

Prognosis in Physical Therapy

Risk of developing a

future pathology

impairment

functional limitation

or disability

ultimate outcome and associated timelines when an impairment or functional limitation is identified.

Results of PT interventions (e.g., goals)

- Functional level

- Need for assistance and/or DME

- Discharge destination

- Return to school, work, leisure

synonyms for factor

indicator, predictor

Prognostic indicator

May predict any type of event or outcome.

Risk factor

Predicts adverse events or outcomes.

Prognosis study designs

Case control study

Cohort study

Predictive model

Case control study

Retrospective comparison of two groups; 1 group with disorder or outcome, 1 group without disorder or outcome.

Case control studies look at

proportion of each group who had the risk factor or prognostic indicator of interest

Cohort study

Prospective comparison of two or more groups before they have the disorder or outcome.

Cohort studies monitor

groups to see who develops the disorder or outcome and identify what characteristics they have

Predictive Model

Can use retrospective or prospective data; regression model determines which relevant factors predict the outcome of interest.

Assessment of Study Credibility

-Does the study of the given test compare to a reference test? (gold standard)

-Were practioners blinded to subject grouping?

-Are important details adequately described?

-Are measures of reliability discussed?

-Are measures of Validity discussed?

Did investigators operationally define their sample?

Addresses the issue of sample homogeneity

Are the subjects representative of the population from which they are drawn?

Addresses the issue of bias in the sample

Did all subjects enter the study at the same (preferably early) stage of their condition?

Addresses the issue of timing for the development of the outcome of interest

inception cohort

subjects enrolled in a study before they develop the outcome of interest

Was the study time frame long enough to capture the outcome(s) of interest?

Addresses potential bias in the results due to insufficient time to develop the outcome

Did the investigators collect outcome data from all subjects enrolled in the study?

Addresses potential bias in the results due to attrition (i.e., drop out)

5 and 20 rule

5% drop is low risk

20% is a threat to valifity

Were outcome criteria operationally defined?

Addresses the methods by which investigators will identify which subjects achieved the outcome of interest

Were the individuals collecting the outcome measures masked to the status of prognostic factors in each subject?

Addresses investigator behavior that may create bias in the sample

Did the sample include subgroups of patients for whom prognostic indicators will differ?

If yes, did the investigators conduct separate subgroup analyses or statistically adjust for these different prognostic factors?

Addresses the possibility that other characteristics of some patients will influence the outcome of interest

Did the authors repeat the study on another set of subjects?

Addresses the reproducibility of the results

Odds Ratios (OR)

The 'odds' that an individual with the prognostic or risk factor will develop the outcome of interest.

Odds ratios are calculated via

logistic regression [Exp(B)]

- sign in front of beta means

inverse relationship between predictor and outcome

Odds ratio formula

(a/b)/(c/d) or (ad)/(bc)

Odds ratio interpretation (negative outcome)

OR > 1 means the odds are in favor of an adverse outcome

OR < 1 means the odds are against an adverse outcome.

Odds ratio interpretation (positive outcome)

OR > 1 means the odds are in favor of a positive outcome

OR < 1 means the odds are against a positive outcome.

Risk ratio (RR) or relative risk

Ratio of risk for developing the adverse outcome in patients with the risk factor versus patients without the risk factor.

Risk ratio depends on

incidence of outcome in both groups

risk ratios can't be used in

case control designs

Risk Ratio formula

((a/a+b)/(c/(c+d))

Risk Ratio interpretation

RR > 1 means there is an increased risk of an adverse outcome

RR < 1 means there is a decreased risk of an adverse outcome.

Hazard ratios

Measure of how often an event happens in one group compared to another

Hazard ratios are shown in conjunction with

survival curves

p-value

The probability that the result (e.g., correlation coefficient, difference, OR, RR) occurred due to chance.

95% Confidence Interval (C.I.)

The range of values within which the true value is estimated to lie within a 95% probability.

If an OR = 1.0

it represents a 50:50 chance of increasing or decreasing the odds that the outcome will occur

If an RR = 1.0

it represents a 50:50 chance increasing or decreasing the risk that the outcome will occur

Should you use this evidence

Is the study high quality (e.g., does the design minimize bias)?

Are the results important enough to use?

Was your patient or client represented in the study sample?

Patient's or client's values and preferences re:

ultimate outcomes

Addresses the natural course of a condition.

individual level prognosis

Future health status

Response to intervention

Duration of treatment

Important for family members, insurance, employers and pt's

population level prognosis

Intervention trials

Development of Clinical guidelines

Prognosis always have an

element of uncertainty

Elements of prognosis

1. The outcome(s) that are possible, 2. The likelihood that the outcome(s) will occur, 3. The time frame required for their achievement.

Risk management

The goal is to prevent adverse events that may occur sometime in the future.

secondary prevention

Treat disease or conditions after they've been diagnosed

tertiary prevention

Manage longer term more complex problems

Diabetes, heart disease, chronic LBP

primary prevention

Measures to prevent disease or impairments from occurring

Fall prevention, injury workplace aversion

Prognostic Factors

May influence the likelihood of an outcome.

Demographic traits

Age, gender, occupation.

Disease-specific factors

Stage, severity, natural history.

Comorbid conditions

CVD, obesity, fear-avoidance beliefs, depression.

Medical comorbidities

Arthritis, Diabetes, CVD, Obesity, Cognitive impairment.

Behavioral comorbidities

Depression, Fear-avoidance.

Prognostic factors do not need to CAUSE the outcome; just need to be

associated to be predictive of the likelihood of the outcome

Study Credibility

Evidence should be evaluated with assessment of research validity.

Higher validity provides greater confidence that findings are

reasonably free from bias

Sample definition

Did the investigators operationally define the sample in their study?

Ensure subjects are individuals who have, or who are at risk for, the outcome of interest

Look for clear operational definition of the disorder

Subject representativeness

Were the subjects representative of the population from which they were drawn?

Sampling bias

exists when a sample is not representative of the population from which it was drawn

referral bias

The systematic error due to the inconsistent recruitment of subjects based on the identification of subjects by clinicians and other study subjects.

Entry into study

Did all subjects enter the study at the same (preferably early) stage of their condition?

inception cohort design

Identifies a group of subjects and follows them forward in time

Preferred starting point of a study about prognostic factors is

right after they become diagnosed or impaired

Length of follow-up

Was the study time frame long enough to capture the outcomes of interest?

Too short of studies will

miss the outcome of interest

Studies of Rare events

Use of case-control design to look retrospectively

Outcomes from all subjects

Did the investigators collect outcome data from all subjects enrolled in the study?

Definition of outcomes

Were outcome criteria operationally defined?

Masking (blinding)

Were individuals collecting outcome measures masked (or blinded) to the status of the prognostic factors for each subject?

tester bias

prior knowledge of prognostic factors introduces expectations that can influence outcomes

Subgroups in sample

Does the sample include subgroups of patients for whom prognostic estimates will differ?

Subgroup-

smaller group that has distinguishing characteristics that separate them from the sample

Confirm findings

Did investigators confirm their findings with a new set of subjects?

Additional Considerations of prognostic studies

Presence or absence of detailed description of the:

Characteristics of sample

Operational definitions, protocols, reliability & validity of measure

Reliability test

Repeating the study on matched people with the same criteria allows for a test of reliability.

Prognostic research uses both

descriptive statistics, as well as tests of relationships to identify prognostic (risk) factors.

Descriptive statistics often reported as

proportions of people who have the risk factor who developed the outcome of interest

A survival curve

Often seen in

epidemiological studies

steeper survival curve slope

worse prognosis