Text Mining and Patient Severity Clusters - Text Mining Techniques for Healthcare Provider Quality Determination - page 296

textmining.diagnoses3digits NOT CONTAINS “v33” and
textmining.diagnoses3digits NOT CONTAINS “v34” and
textmining.diagnoses3digits NOT CONTAINS “v35” and
textmining.diagnoses3digits NOT CONTAINS “v36” and
textmining.diagnoses3digits NOT CONTAINS “v37” and
textmining.diagnoses3digits NOT CONTAINS “v38” and
textmining.diagnoses3digits NOT CONTAINS “v39”
;
Run;

We next examine the ICD9 codes that occur most often in the National Inpatient Sample, omitting
all codes related to childbirth (Table 4).
It is not surprising that hypertension and diabetes are more common than any other patient condi-
tions. Because they are so common, they will not be able to distinguish between different groups of
patients; therefore, some of the less common conditions will be of more value when used to define the
patient clusters (Table 5).
Patients in Cluster 1 have some severe problems largely related to shortness of air and/or pneumonia.
Cluster 2 patients have some severe mental disorders, including drug and alcohol abuse. In contrast,
patients in cluster 4 have diagnoses that are almost exclusively related to drug abuse with no depression.
Those in Cluster 3 have some severe infections, including cellulitis and skin ulcers that are related to
diabetes. Patients in cluster 3 can also have an infection of the blood.
In Cluster 5, patients have problems of the head, neck, and spine. This can also include the condition
of stroke. Cluster 6 patients have kidney failure plus complications that are largely related to dialysis,
including urinary tract infections. In Cluster 7, patients have anxiety plus other chronic problems including
arthritis, shortness of air, and artificial devices. These artifical devices can include orthopedic replace-
ment joints as well as pacemakers. These patients have a need for rehabilitation of some type. Cluster

Table 4. Most common ICD9 codes

Code Translation Frequency Code Translation Frequency

401 Essential hypertension 113,294 285 Other and unspecified anemias 33,153

250 Diabetes 63,615 530 Diseases of esophagus 33,055

276 Disorders of fluid, electrolyte, and acid-base

balance

59,845 599 Other disorders of urethra and urinary

tract

27,347

414 Other forms of chronic ischemic heart

disease

58,221 496 Chronic airway obstruction, not else-

where classified

24,863

272 Disorders of lipoid metabolism 53,956 244 Acquired hypothyroidism 23,097

427 Cardiac dysrhythmias 46,402 715 Osteoarthrosis and allied disorders 22,257

428 Heart failure 43,391 486 Pneumonia, organism unspecified 21,193

305 Nondependent abuse of drug 38,414 493 Asthma 21,170

780 General symptoms 36,393 V15 Other personal history presenting haz-

ards to health

20,438

V45 Other postprocedural states (renal dialysis,

cardiac device in situ)

34,947 518 Other diseases of lung 19,182