Here you will find various documents to support research and teaching, and related to ideas around data archiving, survey data and its use.
Materials have also been put together to support published articles related to some of our data sets and also other projects involving the COMPASS Research Centre. Program code (in SAS) is being made available to allow users to reproduce analyses that were undertaken in the past. You will find these resources among our data holdings as well, as time permits their development and markup. Many thanks in this regard go to our summer scholars for their work on these materials.
New Zealand has been through four national health service market-oriented reforms during the period 1988–2001. Area Health Boards from 1988–1992, Crown Health Enterprises from 1993–1995, Hospital and Health Services 1996–1999 and District Health Boards 2000–current. During this time there was also a substantial reduction in the availability of inpatient beds. The impact of this hospital and system restructuring on the quality and pattern of care is an important issue of public policy concern.
To see if the reduction in public hospital bed availability and multiple reorganisations in New Zealand had an effect on the patterns of care and patient outcomes.
Access to discharge data, amounting to 6,639,487 records, was secured for all 34 major public hospitals in New Zealand over the period 1988–2001. Tables were produced to show the pattern of care over the different regimes. Many different outcomes were looked at including number of discharges, admission rates, access levels, mean length of stay, unplanned readmission rate, and 60-day post-admission mortality rates.
Multilevel models were also used to see the effect the different regimes, as well as other variables, had on 60-day post-admission death, using hospital as a random effect.
Although the number of inpatient beds in use declined by one-third over the period and the national population grew by nearly one-fifth, discharge volumes increased significantly and rates of inpatient admission were maintained, as were access levels for vulnerable groups. These changes were accompanied by workload adjustments (a halving in length of stay and an increase by a quarter in readmission rates). Yet age-adjusted post-admission patient mortality decreased by a quarter over the period of study, a rate of decline that was slowed by the major workload adjustments but not by reform phase.
Other things being equal, a substantial reduction in inpatient bed availability can be effected in national public hospital systems, while largely maintaining access and quality of care. However, the workload adjustments that are required may slow improvements in patient outcomes.
The metadata comprises three sections.
The metadata has been structured in such a way to ensure the data and analytical process is user-friendly and easy to understand. The data dictionary provides a list of all the variables, variable names, a description of variables and their values. It includes both original and derived variables created in the course of the analysis. The annotated SAS code and associated output shows and describes each section of the provided SAS code. More specifically, annotations are given to describe a SAS procedure and the related output. The code is organized by table, according to the journal article. The SAS program contains annotated code without the output. The code is sorted by table and can be downloaded and run in SAS.
We used four data sets in the course of the analysis:
| Data set name | N | Description & filtering |
| h_all_diag | 24,894,974 records | All diagnoses |
| h_all_nhi2 | 6,639,487 records | All patients; applies filtering criteria |
| h_all_last2 | 4,773,663 records | All patients; only the last admission in a given calendar year is kept where there are multiple admissions for a patient in that year |
| patientlast | 3,623,713 records | Inpatients only; only the last admission in a given calendar year is kept where there are multiple admissions for a patient in that year |
What is data filtering
Data filtering is a fundamental and foundational tool of data analysis. It gives you the ability to create a subset of your data based on a set of criteria. Filtering is useful for narrowing your data set to contain only those records pertinent to your analysis, or just narrowing the scope of your analysis, without removing the other observations from the data set.
Data filtering for our data set
The original dataset consisted of 7,682,497 individual records from 34 New Zealand secondary and tertiary hospitals for the period 1988-2001. Filtering was put in place to ensure that definitions of variables were consistent and comparable over time. The fully filtered data set contained 6,639,487 cases.
Below is a sample of the SAS code used for the filtering of the data. It includes filtering of:
Definitions
Boarders: Individuals who stay in hospital without needing treatment, often accompanying another patient, are called 'boarders'. These have been excluded as they are not hospital patients in need of medical care.
Well babies: Healthy hospital-delivered infants were not registered until 1991 and these admissions are not health driven.
SAS Code
data hos1;
set h_all_nhi;
rep = 0
keep event_id nhi evstdate evendate rep;
run;
proc sort data=hos1;
by nhi evstdate descending evendate;
run;
*this code creates a new variable rep that takes the value 1 if both nhi and evstdate are the same;
proc iml;
use hos1;
read all var{event_id nhi evstdate rep};
N = nrow(nhi);
DO i=2 TO N;
if ((nhi[i] = nhi[i-1]) & (evstdate[i]=evstdate[i-1])) then rep[i]=1;
else rep[i]=0;
END;
create hos2 var{event_id rep};
append;
close hos2;
quit;
proc sort data=h_all_nhi; by event_id; run;
proc sort data=hos2; by event_id; run;
proc datasets;
delete hos1;
run;
data h_all_nhi;
merge h_all_nhi hos2;
by event_id;
run;
*this code involves changing the date to year to enable the data to be analysed by year;
data h_all_nhi;
set h_all_nhi;
if "01JAN1988"D <= evendate <= "31DEC1988"D then year = 1988;
if "01JAN1989"D <= evendate <= "31DEC1989"D then year = 1989;
if "01JAN1990"D <= evendate <= "31DEC1990"D then year = 1990;
if "01JAN1991"D <= evendate <= "31DEC1991"D then year = 1991;
if "01JAN1992"D <= evendate <= "31DEC1992"D then year = 1992;
if "01JAN1993"D <= evendate <= "31DEC1993"D then year = 1993;
if "01JAN1994"D <= evendate <= "31DEC1994"D then year = 1994;
if "01JAN1995"D <= evendate <= "31DEC1995"D then year = 1995;
if "01JAN1996"D <= evendate <= "31DEC1996"D then year = 1996;
if "01JAN1997"D <= evendate <= "31DEC1997"D then year = 1997;
if "01JAN1998"D <= evendate <= "31DEC1998"D then year = 1998;
if "01JAN1999"D <= evendate <= "31DEC1999"D then year = 1999;
if "01JAN2000"D <= evendate <= "31DEC2000"D then year = 2000;
if "01JAN2001"D <= evendate <= "31DEC2001"D then year = 2001;
age=(evstdate-dob)/365.25;
if "01JAN1992"D <= evendate <= "31MAY1997"D and facility = 3213 and (650<=drg_31<=727
or specialty='P01' or specialty='P07' or specialty='P10' or specialty='P15'
or specialty='P17' or specialty='P23' or specialty='P25' or specialty='P27'
or specialty='P35' or specialty='P40' or specialty='P42' or specialty='P43'
or specialty='S30' or specialty='S32' or specialty='S33' or specialty='M69'
or 0 <= age <= 0.25) then status='NW';
else status='Ot';
if "01JAN1992"D <= evendate <= "31MAY1997"D and facility = 3213 and
status='Ot' then facility = 3212;
drop status;
if ((evendate=evstdate) & (end_typ ne 'DD') & (substr(specialty,1,3) in
('M05','M06','M07','M08'))) then ae = 1;
else ae = 0;
if admission_source_code='T' then delete; *delete all the transfer patients;
if drg_31=572 then delete;
if drg_31=940 then delete;
if 951<=drg_31<=952 or 955<=drg_31<=956 then delete;
if rep=1 then delete; *delete all patients with the same nhi code and same admission date;
if ae=1 then delete; *delete all A&E cases;
run;
data diag;
set ''; *load the combined diagnosis data file to find well-baby cases;
if ('V30' le substr(clin_cd,1,3) le 'V39');
run;
proc sort data=diag;
by event_id;
run;
proc transpose data=diag out=diag1 prefix=diagcode;
by event_id;
var clin_cd;
run;
data diag1;
set diag1;
wb = 1;
keep event_id wb;
run;
proc sort data=h_all_nhi; by event_id; run;
data h_all_nhi;
merge h_all_nhi diag1;
by event_id;
if wb=. then wb = 0;
run;
data diag;
set '';
bor = 1;
if (diag_typ = 'A') & (substr(clin_cd,1,4) = 'V650');
keep event_id bor;
run;
proc sort data=diag; by event_id; run;
data h_all_nhi;
merge h_all_nhi diag;
by event_id;
if bor=. then bor = 0;
run;
data h_all_nhi;
set h_all_nhi;
if year = . then delete;
if wb = 1 then delete; * delete all well babies cases;
if bor = 1 then delete; *delete all boarder cases;
run;
| SAS procedure | Details |
| PROC UNIVARIATE | Number of beds utilised per year Inpatient average length of stay per year Inpatient average length of stay per year truncated to 97th percentile |
| PROC FREQ | Frequency of discharges per year Frequency of discharges per year only including inpatients Frequency of patients per year Frequency of inpatients per year Percentage of patients who are daystay patients per year Percentage of inpatients admissions which are emergencies per year Frequency and percentage of readmissions per year for all dischargers Frequency and percentage of readmission inpatients per year for all dischargers Frequency and percentage of readmissions per year for all patients Frequency and percentage of readmission inpatients per year for all patients Frequency and percentage of unplanned readmissions per year for all dischargers Frequency and percentage of unplanned inpatient readmissions per year for all dischargers Frequency and percentage of unplanned readmissions per year for all patients Frequency and percentage of unplanned inpatients readmissions per year for all patients |
| Method | Details |
| PROC FREQ | Frequency of discharges per year for all admissions Frequency of discharges per year for all inpatient admissions Frequency of discharges per year for all patients Frequency of discharges per year for all patients-inpatients Percentage of all admissions that are 75 and over Percentage of all inpatient admissions that are 75 and over Percentage of all admissions that are Māori Percentage of all inpatient admissions that are Māori Percentage of all admissions that are deprived Percentage of all inpatient admissions that are deprived Percentage of all admissions that are ambulatory sensitive Percentage of all inpatient admissions that are ambulatory sensitive Percentage of 60-day post-admission deaths per year for all admissions Percentage of 60-day post-admission deaths per year for all inpatient admissions Percentage of 60-day post-admission deaths per year for all patients Percentage of 60-day post-admission deaths per year for all patients-inpatients Age adjusted percentage of 60-day post-admission deaths per year for all admissions Age adjusted percentage of 60-day post-admission deaths per year for all inpatient admissions Age adjusted percentage of 60-day post-admission deaths per year for all patients Age adjusted percentage of 60-day post-admission deaths per year for all patients – patients |
Table 3
|
Method |
Details |
|
PROC |
Mean length of stay for inpatients |
|
PROC FREQ |
Frequency of dischargers per year |
Table 4
|
Method |
Details |
|
PROC GLIMMIX |
Mixed models where covariates are added one at a time to see their effects on the model. Covariates include year, multiple diagnoses, length of stay, unplanned readmisssions, ambulatory sensitive admissions added, total beds available, reform phase and patient effects. |
|
PROC |
Calculates the variance of the marginal linear predictor for each of the mixed models produced. |
5.1.1 Original variables
DT = Death event (NZHIS use only)
ID = Intended day events
IM = Psychiatric inpatient event (include day patients)
IP = Non-psychiatric inpatient event (includ day patients)
GP= General Practitioner event (NZHIS use only)
OP= Outpatient event (NZHIS use only)
AC = Acute admission
AP = Elective admission to a private hospital
RL = Psychiatric patient returned from leave of more than 10 days
WN= Waiting list
WU= Code not used from 20/8/93
ZA = Arranged admission, ACC covered
ZC = Acute, ACC covered
ZP = Private, ACC covered
ZW= Waiting list, ACC covered
Maximum is 3 days for non psychiatric patients.
N= Did not die within 60 days
0= No
0= Non-Maori
0= Else5.1.2 Derived Variables
Variable Name
Description
Values
los2
Length of stay excluding event leave days.
Days
daystay
A If admission and discharge are the same day,
then patient did not require a bed overnight.
Used to calculate inpatient staydeaths1= Daystay
0= Inpatientemergency
Admission type is acute
1= Acute admission
0= Arranged or elective admissionreadmission_unplanned
Whether readmission within the same year was unplanned
Y= Acute admission
N= Arranged or elective
O= Other, i.e. not readmissoinelderly
Patients aged 75 or over are classed as elderly.
1= Elderly
0= Not elderlyagewtal
A weight variable used to calculate the adjusted deaths
Population age weights
agewtin
A weight variable used to calculate the adjusted deaths for inpatients only
Population age weights
5.2.1 Original Variables
ICD-10-AM coding systems the clinical code falls within.
B= Injury
O= Operation / procedure
of intentional or unintentional injury, underlying cause of death,
operation or procedure performed or pathological nature of a tumor.
ICD-10-AM classification code
5.2.2 Derived Variables
75 years of age from causes considered responsive to prophylactic
or therapeutiic interventions deliverable in ambulatory care settings
0= Not ambulatory sensitive
5.3.1 Original Variables
2= No death
2= Female
2= Non Maori
03= Diseases and disorders of the ear, nose, mouth and throat
04= Diseases and disorders of the respiratory system
05= Diseases and disorders of the circulatory system
06= Diseases and disorders of the digestive system
08= Diseases and disorders of the musculoskeletal and connective tissue
09= Diseases and disorders of the skin, subcutaneous tissue and breast
96= Diseases and disorders of the reproductive system
97= Preganancy, childbirth and puerperium; newborn / neonates
99= Others including ungroupable
2= General
d_ahb88_92= AHB
planned_pct
2= Only one diagnosis5.3.2 Derived Variables
Variable Name
Description
Values
totbedsavailp1000
Total beds available per 1000 people for each year
Beds
r1
Reform phase. 1993-1995 (market orientation)
1= AHB regime
0=CHE, HHS, DHB regimesr2
Reform phase. 1996-1999 (modified market)
1= AHB, CHE regimes
0= HHS, DHB regimesr3
Reform Phase. 2000-2001 (return to regionalization)
1= AHB, CHE, HHS regimes
0= DHB regime
Table 1: Bed supply, patient throughput and pattern of care by reform phase, 1988-2001.
Table 2: Access to hospital care, pattern of admission, and patient outcomes by reform phase, 1988-2001.
Table 3: Rates of change overall and by reform phase for key measures.
Table 4: Multilevel analysis of 60-day postadmission death.
/* Davis, P., Lay-Yee, R., Scott, A., Gauld, R.
Do hospital bed reduction and multiple system reform affect patient mortality?:
A trend and multilevel analysis in New Zealand over the period 1988-2001
(2007) Medical Care, 45 (12), pp. 1186-1194. */
/* Table 1 */
/* Karl Parker 16/12/09 */
libname ss '';
options ls=132;
/* All Dischargers data set */
data z;
length evntlvd2 8;
set ss.h_all_nhi2 (keep = year evstdate evendate evntlvd multstay radm admission_type drg_31);
evntlvd2=evntlvd;
los=evendate-evstdate; *Length of stay including event leave days;
if evntlvd2=. then evntlvd2=0;
los2=los-evntlvd2; *Length of stay excluding event leave days
(this is the length of stay variable that will be used in the following analysis);
*Day stay or inpatient calculation;
if evendate=evstdate then daystay=1;
else daystay=0;
*Emergency variable calculation;
if daystay=0 and admission_type in ('AC','ZC')
then emergency=1;
else emergency=0;
*Unplanned readmission variable creation;
if radm=1 then do;
if admission_type in ('AC','ZC','WU')
then radm_unplanned='Y'; *Acute admission type,(therefore readmission was unplanned);
else radm_unplanned='N'; *Arranged or elective admission type;
end;
else radm_unplanned='O'; *Other, i.e. not readmission;
*Include only data from the following dates, to match Table 1;
if year in ('1988','1990','1994','1998','2001');
run;
/* All Patients data set */
data b;
set ss.h_last_nhi2 (keep= evendate radm admission_type evstdate year);
*Day stay or inpatient calculation;
if evendate=evstdate then daystay=1;
else daystay=0;
*Unplanned readmission variable creation;
if radm=1 then do;
if admission_type in ('AC','ZC','WU')
then radm_unplanned='Y'; *Acute admission type,(therefore readmission was unplanned);
else radm_unplanned='N'; *Arranged or elective admission type;
end;
else radm_unplanned='O'; *Other, i.e. not readmission;
*Include only data from the following dates, to match Table 1;
if year in ('1988','1990','1994','1998','2001');
keep year daystay radm radm_unplanned;
run;
/* Table 1: Output */
*Sort variable for table;
proc sort data=z;by year;run;
/*Average number of beds utilized*/
proc univariate data=z noprint;
by year; *Calculate summary stats for each year;
where daystay=0; *Inpatients only;
var los2; *Length of stay;
output out=losum sum=sum;
run;
*Divide sum by number of days in each year;
data losum2;
set losum;
if year=1988 then aub=sum/366; *leap year;
if year=1990 then aub=sum/365;
if year=1994 then aub=sum/365;
if year=1998 then aub=sum/365;
if year=2001 then aub=sum/365;
*Divide by the number of facilities each year;
if year=1988 then beds=aub/33;
if year=1990 then beds=aub/34;
if year=1994 then beds=aub/34;
if year=1998 then beds=aub/33;
if year=2001 then beds=aub/32;
run;
*Print inpatient average utilised beds per day, for each year;
proc print data=losum2 noobs;title 'verage utilised beds by year';var _ALL_;run;
title ' ';run;
Average utilised beds by year
|
year |
sum |
aub |
beds |
|
1988 |
2943717 |
8042.94 |
243.726 |
|
1990 |
2639812 |
7232.36 |
212.717 |
|
1994 |
2428532 |
6653.51 |
195.692 |
|
1998 |
1832856 |
5021.52 |
152.167 |
|
2001 |
1929012 |
5284.96 |
165.155 |
/*Number of dischargers overall*/
proc freq data=z;
tables year;
run;
The FREQ procedure
|
year |
Frequency |
Percent |
Cumulative |
Cumulative |
|
1988 |
353443 |
15.05 |
353443 |
15.05 |
|
1990 |
389571 |
16.59 |
743014 |
31.64 |
|
1994 |
481821 |
20.51 |
1224835 |
52.15 |
|
1998 |
528961 |
22.52 |
1753796 |
74.67 |
|
2001 |
594842 |
25.33 |
2348638 |
100.00 |
/*Impatient dischargers*/
proc freq data=z;
where daystay=0; *Inpatients only;
tables year;
run;
The FREQ procedure
|
year |
Frequency |
Percent |
Cumulative |
Cumulative |
|
1988 |
328843 |
18.15 |
328843 |
18.15 |
|
1990 |
330734 |
18.26 |
659577 |
36.41 |
|
1994 |
365797 |
20.19 |
1025374 |
56.60 |
|
1998 |
375641 |
20.74 |
1401015 |
77.34 |
|
2001 |
410503 |
22.66 |
1811518 |
100.00 |
/*Number of patients per year*/
proc freq data=b;
tables year;
run;
The FREQ procedure
|
year |
Frequency |
Percent |
Cumulative |
Cumulative |
|
1988 |
271186 |
16.02 |
271186 |
16.02 |
|
1990 |
293711 |
17.35 |
564897 |
33.36 |
|
1994 |
351619 |
20.77 |
916516 |
54.13 |
|
1998 |
369412 |
21.82 |
1285928 |
75.95 |
|
2001 |
407307 |
24.05 |
1693235 |
100.00 |
/*Number of inpatients per year*/
proc freq data=b;
where daystay=0; *Inpatients only;
tables year;
run;
The FREQ procedure
|
year |
Frequency |
Percent |
Cumulative |
Cumulative |
|
1988 |
252381 |
19.32 |
252381 |
19.32 |
|
1990 |
249182 |
19.08 |
501563 |
38.40 |
|
1994 |
266354 |
20.39 |
767917 |
58.79 |
|
1998 |
259345 |
19.85 |
1027262 |
78.64 |
|
2001 |
279003 |
21.36 |
1306265 |
100.00 |
/*Inpatient average length of stay per year*/
proc univariate data=z noprint;
by year;
where daystay=0; *Inpatients only;
var los2; *Length of stay;
output out=ialofs mean=mean; *Outputs the mean for each year;
run;
proc print data=ialofs noobs;var _all_;run;
|
year |
mean |
|
1988 |
8.95174 |
|
1990 |
7.98168 |
|
1994 |
6.63902 |
|
1998 |
4.87928 |
|
2001 |
4.69914 |
/*Inpatient average length of stay truncated*/
proc sort data=z;by drg_31;run;
proc rank data=z groups=100 out=inpat1 (keep = year los2 ranklos drg_31);
by drg_31; *Compute ranks for each diagnostic related group;
where daystay=0; *Inpatients only;
var los2; *Length of stay;
ranks ranklos; *Name length of stay ranks;
run;
proc univariate data=inpat1 noprint;
title 'inpatient discharges: los by drg';
by drg_31; *Compute summary stats for each diagnostic related group;
where ranklos=97; *Include 97th percentile only;
var los2; *Length of stay;
output out=inpat2 max=max97; *Output max length of stay (truncated);
run;
data inpat3;
merge inpat1 (in=a) inpat2 (in=b); *merge data sets;
by drg_31;
if a;
if ranklos>97 then los2=max97;
*If length of stay is ranked higher than the 97th percentile, values are assigned the 97th percentile value (max value);
run;
*Sort for summary stats;
proc sort data=inpat3;by year;run;
*Summary stats for inpatient average length of stay truncated to 97th percentile for each year;
proc univariate data=inpat3 noprint;
title 'inpatient discharges: truncated los';
by year; *Compute summary stats for each year;
var los2; *Length of stay;
output out=inpat4 mean=sum; *output sums and name output data set losum;
run;
proc print data=inpat4;title 'Average length of stay by year (truncated)';var _all_;run;
title ' ';run;
Average length of stay by year (truncated)
|
Obs |
year |
average |
|
1 |
1988 |
7.23786 |
|
2 |
1990 |
6.75520 |
|
3 |
1994 |
5.83490 |
|
4 |
1998 |
4.70917 |
|
5 |
2001 |
4.52982 |
proc sort data=z;by year; run;
/*Daystay %*/
proc freq data=z;
by year; *Frequency table calculated for each year;
tables daystay;
run;
The FREQ procedure
year=1988
|
daystay |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
328843 |
93.04 |
328843 |
93.04 |
|
1 |
24600 |
6.96 |
353443 |
100.00 |
year=1990
|
daystay |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
330734 |
84.90 |
330734 |
84.90 |
|
1 |
58837 |
15.10 |
389571 |
100.00 |
year=1994
|
daystay |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
365797 |
75.92 |
365797 |
75.92 |
|
1 |
116024 |
24.08 |
481821 |
100.00 |
year=1998
|
daystay |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
375641 |
71.01 |
375641 |
71.01 |
|
1 |
153320 |
28.99 |
528961 |
100.00 |
year=2001
|
daystay |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
410503 |
69.01 |
410503 |
69.01 |
|
1 |
184339 |
30.99 |
594842 |
100.00 |
/*Energency %*/
proc freq data=z;
by year; *Frequency table calculated for each year;
where daystay=0; *inpatient admissions;
tables emergency; *acute admission;
run;
The FREQ procedure
year=1988
|
emergency |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
120642 |
36.69 |
120642 |
36.69 |
|
1 |
208201 |
63.31 |
328843 |
100.00 |
year=1990
|
emergency |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
118131 |
35.72 |
118131 |
35.72 |
|
1 |
212603 |
64.28 |
330734 |
100.00 |
year=1994
|
emergency |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
133207 |
36.42 |
133207 |
36.42 |
|
1 |
232590 |
63.58 |
365797 |
100.00 |
year=1998
|
emergency |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
126741 |
33.74 |
126741 |
33.74 |
|
1 |
248900 |
66.26 |
375641 |
100.00 |
year=2001
|
emergency |
Frequency |
Percent |
Cumulative |
Cumulative |
|
0 |
130178 |
31.71 |
130178 |
31.71 |
|
1 |
280325 |
68.29 |
410503 |
100.00 |
proc freq data=z;
tables radm*year;
run;
The FREQ procedure
|
Table of RADM by year |
||||||
|
RADM |
year |
|||||
|
Frequency |
1988 |
1990 |
1994 |
1998 |
2001 |
Total |
|
0 |
326014 |
356047 |
436707 |
477413 |
534794 |
2130975 |
|
1 |
27429 |
33524 |
45114 |
51548 |
60048 |
217663 |
|
Total |
353443 |
389571 |
481821 |
528961 |
594842 |
2348638 |
/*Number and percentage of readmission inpatients per year for all dischargers*/
proc freq data=z;
where daystay=0; *Inpatients only;
tables radm*year;
run;
The FREQ procedure
|
Table of RADM by year |
||||||
|
RADM |
year |
|||||
|
Frequency |
1988 |
1990 |
1994 |
1998 |
2001 |
Total |
|
0 |
303249 |
301698 |
332095 |
339441 |
371095 |
1647578 |
|
1 |
25594 |
29036 |
33702 |
36200 |
39408 |
163940 |
|
Total |
328843 |
330734 |
365797 |
375641 |
410503 |
1811518 |
/*Number and percentage of readmissions per year for all patients*/
proc freq data=b;
tables radm*year;
run;
The FREQ procedure
|
Table of RADM by year |
||||||
|
RADM |
year |
|||||
|
Frequency |
1988 |
1990 |
1994 |
1998 |
2001 |
Total |
|
0 |
254937 |
274556 |
328629 |
344315 |
379289 |
1581726 |
|
1 |
16249 |
19155 |
22990 |
25097 |
28018 |
111509 |
|
Total |
271186 |
293711 |
351619 |
369412 |
407307 |
1693235 |
/*Number and percentage of readmission inpatients per year for all patients*/
proc freq data=b;
where daystay=0; *Inpatients only;
tables radm*year;
run;
The FREQ procedure
|
Table of RADM by year |
||||||
|
RADM |
year |
|||||
|
Frequency |
1988 |
1990 |
1994 |
1998 |
2001 |
Total |
|
0 |
236990 |
231930 |
247482 |
239849 |
257950 |
1214201 |
|
1 |
15391 |
17252 |
18872 |
19496 |
21053 |
92064 |
|
Total |
252381 |
249182 |
266354 |
259345 |
279003 |
1306265 |
/*Number and percentage of unplanned readmissions per year for all dischargers*/
proc freq