Transcript Healthcare Statistics

Healthcare
Statistics,
Research and
Epidemiology
©2010 Jones and Bartlett Publishers
Healthcare Statistics
©2010 Jones and Bartlett Publishers
Define Statistics
•
The overall science of extracting information from a
group of data and using the information to make
inference about that larger group of data.
•
The study of variation
•
Descriptive Statistics
–
•
Statistical data collected concerning the attributes of a
population
Inferential Statistics
–
Statistical data collected from a sample to make inference
about the population from which the sample is extracted
©2010 Jones and Bartlett Publishers
Importance of Maintaining
Healthcare Statistics
•
Strategic planning
•
To compare past with current
performance indicators
•
For accreditation compliance
•
Healthcare Agencies
©2010 Jones and Bartlett Publishers
HIM role in Healthcare Statistics
•
Decide if health information collected meets statistical needs of
health care facility
•
Be aware or sources of data within the facility
•
Be prepared to merge other data with data from the health record
•
Collect quality health data
•
Organize data into databases
•
Statistically analyze collected data
•
Develop, generate and interpret health care statistical reports
©2010 Jones and Bartlett Publishers
Definitions
•
Census
–
•
Number of patients present at any given time
Daily Inpatient Census
–
•
Number of patients present at the official census taking time
(usually 12 midnight) each day plus the number of patients
admitted and discharged the same day
Inpatient Service Day
–
•
Unit of measure denoting the services received by one
inpatient in one 24 hour period
Total Inpatient Service Days
–
Sum of all inpatient service days for each of the days in the
period under consideration
©2010 Jones and Bartlett Publishers
Definitions
•
Length of Stay
–
•
The number of calendar days from admission to discharge
Total Length of Stay
–
•
The sum of days stay of any group of inpatients discharged
during a specific period of time
Inpatient Bed Count
–
•
The number of available facility inpatient beds both occupied
and vacant on a given day
Inpatient Bed Count Day
–
The unit of measure denoting the presence of one inpatient
bed either occupied or vacant set up and staffed for use in
one 24 hour period
©2010 Jones and Bartlett Publishers
Definitions
•
Hospital Inpatient Autopsy
–
•
The postmortem examination performed in a hospital facility
(performed by a pathologist or other responsible physician) on
the body of a patient who died during inpatient hospitalization
Hospital Autopsy
–
•
The postmortem examination, wherever performed (by a
pathologist or responsible physician) of the body of a person
who has at some time been a hospital patient
Nosocomial Infection
–
•
Those acquired during hospitalization
Inpatient Discharge Analysis
–
Health record information is reviewed to determine types of services
provided, length of stay, discharge status and other data to assist in
calculating hospital performance indicators.
©2010 Jones and Bartlett Publishers
Definitions
•
Vital Statistics
–
–
•
Crucial events in life such as births, deaths, adoptions,
marriages and divorces
National Center for Health Statistics (NCHS) recommends
standard forms which most states adopt to develop birth,
death and fetal death certificates
Estimate
–
•
Any statistic calculated on a sample of observations
Sampling error
–
The principle that the characteristics of a sample are not
identical to the characteristics of the population from which the
sample is drawn
©2010 Jones and Bartlett Publishers
Definitions
•
Population
–
–
•
Any defined aggregate of objects, persons or events (sum
total)
The variables used as the basis for classification or
measurement being specified
Parameter
–
•
Statistic calculated on a population value
Sample
–
–
Any sub-aggregate drawn from the population
A small group that observed that is drawn from the population
in order to make inference back to the population from which it
was drawn
©2010 Jones and Bartlett Publishers
Common Hospital Services
•
Internal Medical
• Pediatrics
•
Surgery
• Radiology
•
Obstetrics and
Gynecology
• Diagnostic imaging
•
•
• Neurology
Neonatal
• Psychiatry
Anesthesiology
• Pathology
©2010 Jones and Bartlett Publishers
Presentation of Data
• Data should be presented in such a
manner which catches the reader’s
attention, encourages interest and makes
data easy to interpret and use
©2010 Jones and Bartlett Publishers
Tables
•
•
•
Columns of figures, each
labeled to identify
contents
Include title, date, and
person who prepared
table
Gender Age
LOS # Dis
Male
25.6
8.3
85
Female 24.0
4.2
43
Include narrative
explanation of what table
depicts
©2010 Jones and Bartlett Publishers
Graphs
•
Horizontal axis
(independent variable)
•
Vertical axis
(dependent variable)
•
Types of Graphs
–
Bar
•
Used to report count
values of categorical
data
©2010 Jones and Bartlett Publishers
Histogram
•
Graphic
representations
of frequency
distributions
©2010 Jones and Bartlett Publishers
Line
•
Used to
provide a
simple visual
method of
monitoring
trends over
time
©2010 Jones and Bartlett Publishers
Pie Chart
•
Displays
frequencies
in each
category
©2010 Jones and Bartlett Publishers
Pareto Chart
• Type of bar
graph which
displays
categories of
data in
descending
order of
frequency or
significance
16
120.00%
14
100.00%
12
80.00%
10
8
60.00%
Frequency
Cumulative
6
40.00%
4
20.00%
2
0
0.00%
Dislike
Neutral
Very Good
Excellent
©2010 Jones and Bartlett Publishers
Scatter diagram
•
Used to plot
the points for
two variables
that may be
related to
each other
©2010 Jones and Bartlett Publishers
Frequency polygon
•
Similar to a
histogram
•
Graph of a
frequency
distribution in
line form
rather than a
bar graph
©2010 Jones and Bartlett Publishers
Ratios, Proportions, and Rates
• General formula for calculating Ratios,
Proportions and Rates:
• Ratio, proportion, rate = x/y x 10n
©2010 Jones and Bartlett Publishers
Ratios, Proportions, and Rates (cont.)
•
Ratios
–
•
The quantities being compared may be expressed so that x
and y are completely independent of each other, or x may be
included in y
Proportions
–
–
•
A type of ratio in which x is a portion of the whole (x + y)
The numerator is always included in the denominator
Rates
–
Used to measure events over a period of time
©2010 Jones and Bartlett Publishers
Measures of Central Tendency
•
Mean
–
•
Average calculated by adding the values of all observations
and dividing the total by the number of observations
Median
–
•
Middle most value when values are ranked in numeric order
Mode
–
–
–
Value that occurs most frequently
When no value repeats more than once, there is no mode
When several values repeat with the same frequency, each is
the mode
©2010 Jones and Bartlett Publishers
Measures of Variability
•
Measures of Dispersion, the amount of
variability of the measurement around
the mean or median. The degree to
which numerical data tend to be spread
about an average value
©2010 Jones and Bartlett Publishers
Measures of Variability
•
Range
–
•
The difference between the highest and lowest values
Variance
–
–
•
Demonstrate how values are spread or dispersed around the mean
Computed by squaring each deviation from the mean, summing them
and then dividing their sum by the degrees of freedom (n-1)
Standard Deviation
–
–
–
–
–
Demonstrate how values are spread or dispersed around the mean
The most common measure of variation
The square root of the variance
Small standard deviation demonstrates data are close to mean and a
large standard deviation means data are more spread out from the
mean
Example: December discharges for Houston Hospital had a mean of
6 days and a standard deviation of 2. Therefore, if a patient stayed
in the hospital one standard deviation above the mean, he had a
length of stay of 8 days (6 + 2 = 8). If a patient had a length of stay
of 2 days, then he was in the hospital 2 standard deviations below
the mean (6 – 2 – 2 = 4)
©2010 Jones and Bartlett Publishers
Normal Distribution
•
Measures of central tendency and variation are
interpreted as they relate to the normal distribution
•
Theoretical family of distributions that may have any
mean or any standard deviation
•
A bell-shaped curve (also referred to as “Normal
Curve”) that is symmetrical about the mean
–
50% of observations fall above the mean and 50% fall below
the mean
Each side of the mean extends to a tail
–
•
•
When the research hypothesis is directed to only one end of the
curve, it is considered a one-tailed test
When the research hypothesis is directed to both ends of the
curve, it is considered a two- tailed test
©2010 Jones and Bartlett Publishers
Normal Curve
Mean
Median
Mode
=0
©2010 Jones and Bartlett Publishers
Hospital Performance Indicators
•
•
•
•
Number of Admissions
and Discharges
•
Mortality rates
•
Autopsy rate
•
Infection rates
•
Consultation rate
•
Other indicators of
hospital performance is
requested or required
Average Daily Census
Average Length of Stay
Occupancy rate
©2010 Jones and Bartlett Publishers
Average Daily Census (formula)
Total inpatient service days for a period
(excluding newborn)
---------------------------------------------------------Total number of days in period
©2010 Jones and Bartlett Publishers
Average Daily Census
(calculated)
4,518
-----------------31
= 145.74
= 146
May, 200x
Unit
Inpatient Service Days
Medicine
1,656
Surgery
1,987
OB
875
TOTAL
4,518
©2010 Jones and Bartlett Publishers
Average Length of Stay (formula)
Total length of stay or discharge days
(excluding newborns)
---------------------------------------------------Total discharges
©2010 Jones and Bartlett Publishers
Average LOS (calculated)
1055
------------------ =
673
1.56 = 1.6
# admissions
610
# discharges
673
# pts remaining last day of
month
# discharge days
198
1055
# inpatient service days
1113
# of beds
301
©2010 Jones and Bartlett Publishers
Occupancy rate (formula)
Total inpatient service days for a period x 100
----------------------------------------------------------------Total inpatient bed count days in period under
consideration (beds x days)
©2010 Jones and Bartlett Publishers
Occupancy rate (calculated)
1113
------------------ x 100 =
301 x 30
12.3%
June, 200x
# admissions
610
# discharges
673
# pts remaining last day of
month
# discharge days
198
1055
# inpatient service days
1113
# of beds
301
©2010 Jones and Bartlett Publishers
Death rate (formula)
Total # deaths (including newborns) for a
period x 100
--------------------------------------------------------Total # discharges (including deaths)
Also referred to as
Gross Death Rate
©2010 Jones and Bartlett Publishers
Death rate (calculated)
(7 + 3) x 100
------------------
688 + 90
= 1.29%
# admissions
685
# discharges
688
# deaths
# newborn discharges
#newborn deaths
7
90
3
©2010 Jones and Bartlett Publishers
Post-Operative Death rate
(formula)
Number of post-op deaths for a period x 100
-----------------------------------------------------Number of patients operated upon
©2010 Jones and Bartlett Publishers
Post-op Death rate (calculated)
2 x 100
----------------
10,111 + 1,523
= .02%
# patients operated upon
10,111
# newborn pts operated upon 1,523
# adult/children deaths
41
# newborn deaths
2
# deaths within 10 days post-op
2
©2010 Jones and Bartlett Publishers
Fetal Death rate (formula)
Number of intermediate and late fetal
deaths for a period x 100
-----------------------------------------------Total number of live births +
intermediate and late fetal deaths for the
period
©2010 Jones and Bartlett Publishers
Fetal Death rate (calculated)
(2 + 4) x 100
-----------------------
= 4.58%
125 + 2 + 4
During January, 200x, Houston hospital had 125 live births, 2
intermediate fetal death, and 4 late fetal deaths. What was the
fetal death rate for January?
©2010 Jones and Bartlett Publishers
Maternal Death rate (formula)
Number of direct maternal deaths
for a period x 100
------------------------------------------
Number of OB discharges
(including deaths) for the period
©2010 Jones and Bartlett Publishers
Maternal Death rate (calculated)
3 x 100
-----------------
= .21%
1411
During March, 200x, there were 3 maternal deaths following
c-sections. In addition, 5 patients had abortions. The OB unit
admitted 1405 patients and discharged 1411 patients. What was the
maternal death rate for the hospital?
©2010 Jones and Bartlett Publishers
Gross Autopsy rate (formula)
Total inpatient autopsies for a period x 100
----------------------------------------------------------Total inpatient deaths for the period
©2010 Jones and Bartlett Publishers
Gross Autopsy rate (calculated)
10 x 100
------------------------
= 32.26%
31
During December, 200x, there were 1001 discharges, 31 deaths (including
newborns), and 10 autopsies. What was the gross autopsy rate?
©2010 Jones and Bartlett Publishers
Net Autopsy rate
Total inpatient autopsies x 100
-------------------------------------------------Total inpatient deaths – unautopsied
coroners’ or medical examiners’ cases
©2010 Jones and Bartlett Publishers
Net Autopsy rate (calculated)
5 x 100
--------------------
= 71.43%
10 - 3
During April 200x, there were 559 discharges, 10 inpatient deaths
and 5 autopsies. Three (3) deaths were unavailable for autopsy
because they were released to the coroner. What was the Net
Autopsy Rate for the month?
©2010 Jones and Bartlett Publishers
Adjusted Hospital Autopsy rate
(formula)
Total hospital autopsies x 100
-------------------------------------------------Total number of deaths of hospital patients
whose bodies are available for hospital
autopsy
©2010 Jones and Bartlett Publishers
Adjusted Hospital Autopsy rate
(calculated)
5 x 100
---------------
= 16.13%
36 + 2 - 4 - 3
During September 200x, 36 inpatient deaths occurred. Two (2)
outpatients died and their bodies were brought to the hospital.
Among these, 4 deaths were reported to the coroner, 3 were
transferred to another city therefore no autopsy was performed,
and 5 hospital autopsies were performed. What was the Adjusted
Hospital Autopsy Rate?
©2010 Jones and Bartlett Publishers
Post-operative Infection rate
Number of infections in clean surgical
cases for a period x 100
------------------------------------------------Number of surgical operation for the
period
©2010 Jones and Bartlett Publishers
Post-Op Infection rate (calculated)
3 X 100
-----------
= .37%
802
During August 200x, 802 surgical operations were performed. The
infection control committee reported 3 post-operative infections in clean
surgical cases. What was the Post-Operative Infection Rate?
©2010 Jones and Bartlett Publishers
Calculating Any Percentage
What you actually have
---------------------------------------- x 100
What you could have had
©2010 Jones and Bartlett Publishers
Calculate Retrieval Rate
153
----------------- x 100
159
= 96.2%
Clerk retrieved 153 of the 159
charts requested by the
pulmonary clinic. What was
the retrieval rate?
©2010 Jones and Bartlett Publishers
Research
©2010 Jones and Bartlett Publishers
Research
•
Scientific inquiry or question to make
improvements; to increase the body of
knowledge. May be applied of basic.
•
Applied Research
–
•
Improvement of actual practice
Basic
–
Theory building
©2010 Jones and Bartlett Publishers
Research Terms
•
Independent variable
–
–
•
The variable to be manipulated
Also called the experimental or treatment
variable
Dependent variable
–
–
The variable that is measured to determine
the effects of the experimental treatment
Also referred to as the control
©2010 Jones and Bartlett Publishers
Research Terms (cont.)
•
Reliability
–
•
Accuracy of the data in the sense of its
ability to be reproduced or its ability to yield
the same results on repeated trials
Validity
–
–
Degree to which an instrument measures
what it should measure
Assesses relevance, completeness,
accuracy and correctness
©2010 Jones and Bartlett Publishers
Types of Samples
•
Random sample
–
–
•
Every member of the population has an equal
probability of being included
Example: 15 names place in a hat to draw out 10
names
Cluster sample
–
–
–
Random selection of a number of subjects in
naturally occurring groups or clusters
A unit chosen is not an individual but a group of
individuals who are naturally together
Example: Mailing zip code
©2010 Jones and Bartlett Publishers
Types of Samples (cont.)
•
Stratified sample
–
–
•
Sampling of a population which consists of a
number of subgroups or strata that may differ in
characteristics being studied
Example: Ethnic groups
Systematic sample
–
–
Drawing a sample by taking every nth from a list of
the population
Example: List of 100 names, select every 10th
person's name on list
©2010 Jones and Bartlett Publishers
Scientific Method
•
Define the problem (Statement of the Problem)
–
Determine population under study
•
Review the literature
•
Formulate a hypothesis
–
–
Define the null and alternative hypotheses
State the independent and dependent variables
©2010 Jones and Bartlett Publishers
Scientific Method (cont.)
•
Select a research method or design
–
–
–
–
–
–
–
–
–
Experimental
Observational study
Surveys and Questionnaires
Interviews
Historical-prospective
Participant observation
Cross-sectional or Prevalence Study
Cohort study
Case Control
©2010 Jones and Bartlett Publishers
Scientific Method (cont.)
•
Collect the data from sample abstracted from
population
•
Analyze the results
–
–
–
•
Test of significance (t-test, chi-square, ANOVA, etc)
Compare computed result of p-value to alpha level or
compare test statistic to critical value
Accept or reject null hypothesis
Draw conclusions
©2010 Jones and Bartlett Publishers
Test of Significance
•
Purpose is to determine whether observed
differences between groups or relationships
between variables in the sample being studied
are likely to be due to sampling error or are
likely to reflect true differences or relationships
in the population of interest
•
The method utilized to test the null and
alternative hypotheses and to determine
whether or not to accept or reject the null
hypothesis
©2010 Jones and Bartlett Publishers
Test of Significance (cont.)
•
Hypothesis
–
–
•
Identifies the goal of the research and poses a
tentative assumption to be tested
Indicates the nature of the difference or relationship
that is being tested
Null Hypothesis (symbolized as Ho)
–
•
States there is no difference or relationship in the
population under study
Alternative Hypothesis
–
States there is a difference or relationship in the
population under study
©2010 Jones and Bartlett Publishers
Commonly Methods to test
Null Hypothesis
•
T-test
–
–
–
–
–
–
–
Determines if there is a significant difference
between two groups with respect to the
independent and dependent variables
Independent variable
The variable to be manipulated
Also called the experimental or treatment variable
Dependent variable
The variable that is measured to determine the
effects of the experimental treatment
Also referred to as the control
©2010 Jones and Bartlett Publishers
Commonly Methods to test Null Hypothesis (cont.)
•
Chi-square
–
–
•
Determines if there is a significant difference between
observed and expected frequencies
Used for nominal data
Pearson Correlation Coefficient
–
–
–
–
–
–
–
–
Expressed as “r”
Ranges from 0 to ± 1
Used to assess the direction and degree of relationship
between two variables
As r approaches 0 there tends to be less correlation between
the variables; as r approaches 1, there tends to be more
correlation
Coefficient of determination
r-squared (symbolized by r2)
Tells how much of the variation in y is accounted for by the x
variable
If r = .80 then r2 = .64 and 64% of variation in y is accounted
for the x variable
©2010 Jones and Bartlett Publishers
Commonly Methods to test Null Hypothesis (cont.)
•
Regression Analysis
–
–
–
–
–
–
–
Determines the extent one or more explanatory
variables can predict an outcome variable
Coefficient of determination (r2)
Ranges from 0 to 1
Represents the squared correlation between the
explanatory variable(s) and the outcome variable
The value of r2 indicates the proportion of variability
in the outcome that is explained by the predictor
variable(s)
The closer r2 is to one, the stronger the prediction
P value associated with r2 indicates the probability
that the observed value of r2 could occur through
sampling error alone
©2010 Jones and Bartlett Publishers
Commonly Methods to test Null Hypothesis (cont.)
•
ANOVA
–
–
Analysis of Variance
Determines if there is a significant difference
between two or more groups
©2010 Jones and Bartlett Publishers
Test Statistic or t-stat
•
The absolute value of the t-stat is compared to
the critical value. If the t-stat is greater than
the critical value, then the researcher will
reject the null hypothesis.
–
Measures the size of the difference or relationship
observed in the sample.
The probability that the observed value of the test
statistic could occur in the event that the null
hypothesis is true is determined. This is called the
p-value, which ranges from 0 to 1.
–
•
•
P-value answers the question: how likely is it that the
observed difference or relationship is due to chance or due
to sampling error?
As the p-value approaches 0, the smaller the probability
that the observed difference or relationship is due to
chance or sampling error.
©2010 Jones and Bartlett Publishers
Level of Significance
•
Also referred to as alpha level and
symbolized by the Greek letter ά
•
P-value is compared to level of
significance to determine whether to
accept or reject the null hypothesis. If
the p-value is less than the alpha level,
then the researcher will reject the null
hypothesis.
©2010 Jones and Bartlett Publishers
Common Levels of Significance
•
.05 means the decision will be to reject
the null hypothesis if the probability is
smaller than 5 in 100 that the observed
difference or relationship could be due to
sampling error.
•
Setting it at .01 means the decision to
reject will be made if the probability is
smaller than 1 in 100.
©2010 Jones and Bartlett Publishers
Test of Significance Errors
•
Type I error
–
•
Rejecting the null hypothesis when it is true
Type II error
–
Accepting the null hypothesis when it is
false
©2010 Jones and Bartlett Publishers
Epidemiology
©2010 Jones and Bartlett Publishers
Epidemiology
•
The study of disease and the determinants of
disease in populations
•
The study of clinical and health care trend or
patterns and the ability to recognize trends or
patterns with large amounts of data
•
The study of the distribution and determinants
of diseases and injuries in human populations
©2010 Jones and Bartlett Publishers
Common Epidemiological Terms
•
Health
– State of complete physical, mental and social
well being and not merely the absence of
disease
•
Levels of prevention
– Primary – prevention by reducing exposure
– Secondary – early detection and treatment
– Tertiary – alleviation of disability resulting from
disease
•
Rehabilitation
– Attempt to restore an affected individual to a
useful, satisfying and self-sufficient role in
©2010 Jones and Bartlett Publishers
society
Common Epidemiological Terms
•
Risk Factor
– Associated with an increased likelihood that the
disease will develop at a later time
•
Cohort
– A group under study for a period of time
•
Epidemic
– The occurrence in a community or region of a
group of illnesses of similar nature, clearly in
excess of normal expectancy
•
Endemic
– Occurrence that is the habitual presence of a
disease or infectious agent within a geographical
area or the usual prevalence of a given disease
within such area
©2010 Jones and Bartlett Publishers
Common Epidemiological Terms
•
Prevalence Rate (PR)
– The number of existing cases of a disease in a specified
time period divided by the population at that time
– Describes the magnitude of an epidemic
•
Incident rate (IR)
– The number of newly reported cases of a disease in a
specified time period divided by the population at that
time
– Used to compare the frequency of disease in populations
•
Relative risk (RR)
– Used to determine which groups have a greater risk of
developing the disease under study
©2010 Jones and Bartlett Publishers
Epidemiological Research
•
Descriptive Cross Sectional Prevalence Study
– Concurrently describes or examines the distribution
of disease or characteristics and health outcomes
at one specific point or period in time
– Used when little is known about the disease or
characteristic under study
– Used to generate hypotheses, not to test them
•
Case-Control or Retrospective
– Analytical study design in which a disease or health
condition is examined to determine possible causes
– Researcher collects data on disease and controls
by looking back in time
©2010 Jones and Bartlett Publishers
Epidemiological Research (cont.)
•
Prospective
– Determines whether the characteristics or
suspected risk factors preceded the disease or
health condition
•
Cohort
– Prospective study
– Subjects are separated into two groups based
upon their exposures or health characteristics
and then followed forward to determine whether
they develop the disease
©2010 Jones and Bartlett Publishers
Epidemiological Research (cont.)
•
Historical Prospective
– Past records are used to collect information
regarding the exposure characteristics or risk
factors under study
•
Experimental Studies for Clinical and Community
Trials
– Modifies the health characteristics that are
found to cause the disease by using health care
interventions that control progression of the
disease or prevent the disease from occurring
©2010 Jones and Bartlett Publishers