With the background work done in the previous posts, let’s get to the meat of the problem. I’m currently working with a team from an excellent MIT course, Secondary Analysis of Health Records, led by Dr. Leo Anthony Celi. My team is composed of 4 clinicians and 3 data scientists, and we are looking into the problem of “frequent fliers” in the ICU - patients who frequently cycle into the ICU for care. If a classifier can be created to predict which patients are likely to be in this group, a hospital may be able to create a program addressing these patients to help prevent these readmissions.

Overview

An interesting aspect of the question is that many of the factors likely contributing to a patient being a “frequent flier” are not well represented in structured medical records. As an example, a patient who is a substance abuser is very likely to be a “frequent flier”, but many of these patients would not have any specific label identifying them as a substance abuser. Luckily we have access to the notes from the medical staff written during the patient’s stays, and in these notes are often great clues indicating the conditions we’re interested in. Our task, then, is to design an NLP-based classifer that will process these notes and provide an indication of whether any of our concepts are present, e.g. whether the patient is a substance abuser.

Our team is pursuing a supervised learning approach, in which the clinicians on our team have labeled 1000 notes randomly selected from a pool of 40,000, and the data scientists are training classifiers for our concepts. What we’ve found is that some concepts are very low prevalence - e.g. for 1000 notes there are concepts that only have 20 or fewer positive examples. The problem this presents is that the machine learning algorithms are likely to discover spurious correlations that are not true in general application - e.g. if all 20 patients happen to be from a neighborhood that is referenced in the note then the classifier may “think” that a note with that neighborhood referenced means the patient is a substance abuser, and if that neighborhood isn’t referenced then the patient is not a substance abuser.

In order to create a more balanced dataset we decided to enlarge the dataset by having the clinicians label additional notes, but wanted to select notes that were more likely than chance to contain our concepts.

One option was to use our existing text-based classifiers to choose un-annotated notes, annotate those new notes, and use those all to train new classifiers. The problem with this option is that its circular and will re-enforce the existing classifier biases: notes that are in the true-positive region of the classifier will be selected for annotation and training, but the were already being correctly identified, false-positive notes will be selected and annotated which may be helpful to improve specificity, but false-negative notes will not be included and so sensitivity will not improve.

The alternate option I pursue here is to use a separate mode of data for a secondary classifier that will only be used to select notes for annotation. I base this classifier on ICD9 diagnosis codes during the patient’s stay. In this way I avoid the circular problem of using the same classifier to choose its own training data - there still may be issues of bias, but hopefully this will be minimized, and will optimize our annotation time.

Approach

The full notebook is available here, but the main first result of the notebook is highlighted below. This first result is an inspection of how informative the ICD9 codes are to whether the notes will have a particular concept or not.

A summary of the steps taken are as follows:

  1. Load patient data from previous post
  2. Inspect the number of ICD9 codes in the dataset

    This is especially important because of the low prevalence of notes and relatively small overall dataset. If the dataset is much “wider” than it is “tall” - if there are relatively few labels compared to the number of possible features to consider - it is unlikely that resulting classifiers will perform well on general datasets as overfitting is likely. Here we find that there were 1852 raw ICD9 codes found. When we go to only top-level ICD9 codes using the previously explored ICD9 taxonomy we get only 42 codes. Going to one level down we get 181 codes, and two levels down we get 632 codes.

  3. Choose an ICD9 clustering level, assign each note with a 1 or a 0 whether that ICD9 cluster was present

    These will end up being the input feature vectors for our classifier - the presence of ICD9 clusters being used to predict whether the note has a particular class label. I start with “level 1” as by experience 181 codes seemed feasible, but this would just be a first pass - for a real problem we could try different levels and determine the best.

  4. Get an idea of how informative the ICD9 codes will be for building a classifier by looking at the odds ratios. Essentially, looking for codes that are more likely than average to be present for particular concepts. If a concepts has many ICD9 codes that are much more likely to be present than it’s likely we’ll be able to build a good classifer, whereas if all of the codes are present in a distribution similar to the population it will be challenging (though possibly combinations of codes are more likely present, etc.)

Conclusion

It turns out that this worked better than expected! Several concepts have expected ICD9 code relationships, a great indication that this method is given sensible results overall. Some examples below:

  • Advanced.Cancer’

    • (code, 190-199) MALIGNANT NEOPLASM OF OTHER AND UNSPECIFIED SITES
    • (code, V87) OTHER SPECIFIED PERSONAL EXPOSURES AND HISTORY PRESENTING HAZARDS TO HEALTH
    • (code, 170-176) MALIGNANT NEOPLASM OF BONE, CONNECTIVE TISSUE, SKIN, AND BREAST
  • Advanced.Lung.Disease’

    • (code, 980-989) TOXIC EFFECTS OF SUBSTANCES CHIEFLY NONMEDICINAL AS TO SOURCE
    • (code, E869) Accidental poisoning by other gases and vapors
    • (code, 799.1) Respiratory arrest
    • (code, V46) Other dependence on machines and devices
    • (code, 460-466) ACUTE RESPIRATORY INFECTIONS

What was interesting was looking at the more nebulous concepts such as alcohol abuse or depression, concepts that aren’t necessarily well represented by the billing codes. We see that a cluster of ICD9 codes appears to be more likely for these groups in ways that are sensible and expected in hindsight, but surprising that they were found.

  • Alcohol.Abuse’

    • (code, 840-848) SPRAINS AND STRAINS OF JOINTS AND ADJACENT MUSCLES
    • (code, V62) Other psychosocial circumstances
    • (code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted
    • (code, E967) Perpetrator of child and adult abuse
    • (code, V69) Problems related to lifestyle
  • Depression’

    • (code, E931) Other anti-infectives
    • (code, E960) Fight, brawl, rape
    • (code, E967) Perpetrator of child and adult abuse
  • Other.Substance.Abuse’

    • (code, E960) Fight, brawl, rape
    • (code, 840-848) SPRAINS AND STRAINS OF JOINTS AND ADJACENT MUSCLES
    • (code, 070-079) OTHER DISEASES DUE TO VIRUSES AND CHLAMYDIAE

In the next post we’ll follow this up by building classifiers and seeing what kind of performance we can attain.

Notebook

Full notebook available here

Inspect data

In [23]:
note_icd9_df.shape
Out[23]:
(89273, 5)
In [24]:
note_icd9_df.query('level == "source"').shape
Out[24]:
(21044, 5)
In [25]:
grouped = note_icd9_df.query('level == "source"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
code
4280     690
4019     570
42731    499
5849     386
41401    347
Name: md5, dtype: int64
(1852,)
In [26]:
grouped = note_icd9_df.query('level == "0"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
code
390-459    4213
460-519    1825
240-279    1801
580-629    1349
520-579    1223
Name: md5, dtype: int64
(41,)
In [27]:
grouped = note_icd9_df.query('level == "1"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
code
420-429    2085
270-279    1320
510-519     937
580-589     833
996-999     699
Name: md5, dtype: int64
(181,)
In [28]:
grouped = note_icd9_df.query('level == "2"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
code
428    992
276    788
427    755
518    618
401    583
Name: md5, dtype: int64
(632,)
In [29]:
grouped = note_icd9_df.query('level == "3"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
code
428.0    690
401.9    570
427.3    546
518.8    505
414.0    408
Name: md5, dtype: int64
(1017,)
In [30]:
grouped = note_icd9_df.query('level == "4"').groupby('code').count().sort_values('md5', ascending=False)['md5']
display.display(grouped.head())
print(grouped.shape)
Series([], Name: md5, dtype: int64)
(0,)

From above, see that there are no “level 4” codes. As we increase from level 0 to level 3 we get more specific codes, with corresponding increase in number of codes and decrease in the maximum frequency of occurence.

Assemble data for classification

As a first pass, start with a single diagnosis level, combine with labels, inspect

In [27]:
icd9_1lev = note_icd9_df.query('level == "1"')
icd9_1lev.head()
Out[27]:
code level md5 note_row_id subject_id
3 393-398 1 be74552c73a0f9895c4f372763054d26 1414073.0 11590
8 393-398 1 be74552c73a0f9895c4f372763054d26 1414073.0 11590
13 393-398 1 be74552c73a0f9895c4f372763054d26 1414073.0 11590
18 042 1 be74552c73a0f9895c4f372763054d26 1414073.0 11590
22 510-519 1 be74552c73a0f9895c4f372763054d26 1414073.0 11590
In [28]:
labels_df.head()
Out[28]:
subject_id category md5 operator Advanced.Cancer Advanced.Heart.Disease Advanced.Lung.Disease Alcohol.Abuse Chronic.Neurological.Dystrophies Chronic.Pain.Fibromyalgia Dementia Depression Developmental.Delay.Retardation Non.Adherence None Obesity Other.Substance.Abuse Schizophrenia.and.other.Psychiatric.Disorders Unsure row_id_m3 total_m3_distance
0 9973 Discharge 56f2598342cce321539d8975809d487c JTW 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 NaN NaN
1 9973 Discharge 56f2598342cce321539d8975809d487c ETM 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 NaN NaN
2 3365 Discharge eaea5c3c7577135a83f1f0fb583e0d53 JTW 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 17170.0 0.000884
3 3365 Discharge eaea5c3c7577135a83f1f0fb583e0d53 ETM 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 17170.0 0.000884
4 27290 Discharge e7433c0b75ea00346390f029bb830774 JW 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 50828.0 0.001554
In [29]:
tmp_df = icd9_1lev.groupby(['subject_id', 'md5', 'code']).agg({'level': lambda x: 1})
tmp_df.rename(columns={'level': 'code'}, inplace=True)
icd9_vec_df = tmp_df.unstack(fill_value=0)
#icd9_vec_df.columns = icd9_vec_df.columns.droplevel()
In [30]:
icd9_vec_df.head()
Out[30]:
code
code 001-009 030-041 042 047 050-059 062 070-079 110-118 120-129 130-136 137-139 140-149 150-159 160-165 170-176 179-189 190-199 210-229 235-238 239 240-246 249-259 260-269 270-279 290-299 V18 V26 V42 V43 V44 V45 V46 V49 V50 V53 V54 V55 V58 V59 V60 V62 V63 V64 V65 V66 V69 V70 V85 V87 V88
subject_id md5
68 27572b36bd4c26c322f50cf65d095d16 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
109 27d1f5907fa14b6702837a845f84c54e 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
3e0fff775cfb678fdfa06ece68ebfab5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
8efc0a2ff698b75ce183e3183c1bf204 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
f5f69772c32f1b0ac05b7cf408f7a6db 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

5 rows × 181 columns

In [125]:
feat_vecs = labels_df.groupby(['subject_id', 'md5']).max()[['category',] + categories]
feat_vecs = feat_vecs.join(icd9_vec_df)
/mnt/cbds_homes/ecarlson/.local/lib/python3.4/site-packages/pandas/tools/merge.py:480: UserWarning: merging between different levels can give an unintended result (1 levels on the left, 2 on the right)
  warnings.warn(msg, UserWarning)
In [126]:
feat_vecs.head()
Out[126]:
category Advanced.Cancer Advanced.Heart.Disease Advanced.Lung.Disease Alcohol.Abuse Chronic.Neurological.Dystrophies Chronic.Pain.Fibromyalgia Dementia Depression Developmental.Delay.Retardation Non.Adherence None Obesity Other.Substance.Abuse Schizophrenia.and.other.Psychiatric.Disorders Unsure (code, 001-009) (code, 030-041) (code, 042) (code, 047) (code, 050-059) (code, 062) (code, 070-079) (code, 110-118) (code, 120-129) (code, V18) (code, V26) (code, V42) (code, V43) (code, V44) (code, V45) (code, V46) (code, V49) (code, V50) (code, V53) (code, V54) (code, V55) (code, V58) (code, V59) (code, V60) (code, V62) (code, V63) (code, V64) (code, V65) (code, V66) (code, V69) (code, V70) (code, V85) (code, V87) (code, V88)
subject_id md5
68 27572b36bd4c26c322f50cf65d095d16 Nursing/Other 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
109 27d1f5907fa14b6702837a845f84c54e Nursing/Other 0 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
3e0fff775cfb678fdfa06ece68ebfab5 Nursing/Other 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
8efc0a2ff698b75ce183e3183c1bf204 Nursing/Other 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
f5f69772c32f1b0ac05b7cf408f7a6db Discharge 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

5 rows × 197 columns

In [132]:
feat_vecs.shape
Out[132]:
(1849, 197)
In [127]:
feature_cols = [c for c in feat_vecs.columns if isinstance(c, tuple)]
In [128]:
code_lookup = [{'icd9':icd, 'descr': tree.find(icd[1]).description} for icd in feature_cols]
In [129]:
code_lookup_df = pd.DataFrame(code_lookup).set_index('icd9')

Note: not really an odds ratio, dividing by population mean rather than mean of non-flagged population, otherwise many divide by zeros

In [130]:
all_vecs = feat_vecs[feature_cols].mean()
likely_concepts = dict()
for cat in categories:
    with_label = feat_vecs.loc[feat_vecs[cat]==1, feature_cols].mean()
    with_label = with_label/all_vecs
#     no_label = feat_vecs.loc[feat_vecs[cat]==0, feature_cols].mean()
#     with_label = with_label/no_label
    with_label.name = 'OR'
    with_label = code_lookup_df.join(pd.DataFrame(with_label))
    likely_concepts[cat] = with_label.sort_values('OR', ascending=False)
In [131]:
for cat in categories[:15]:
    print(cat)    
    display.display(likely_concepts[cat].head(20))
Advanced.Cancer
descr OR
icd9
(code, 190-199) MALIGNANT NEOPLASM OF OTHER AND UNSPECIFIED SITES 12.513294
(code, V87) OTHER SPECIFIED PERSONAL EXPOSURES AND HISTORY PRESENTING HAZARDS TO HEALTH 9.144330
(code, E933) Primarily systemic agents 7.837997
(code, 170-176) MALIGNANT NEOPLASM OF BONE, CONNECTIVE TISSUE, SKIN, AND BREAST 6.858247
(code, E853) Accidental poisoning by tranquilizers 6.096220
(code, V14) Personal history of allergy to medicinal agents 6.096220
(code, 795) Other and nonspecific abnormal cytological, histological, immunological and DNA test findings 6.096220
(code, 140-149) MALIGNANT NEOPLASM OF LIP, ORAL CAVITY, AND PHARYNX 6.096220
(code, 150-159) MALIGNANT NEOPLASM OF DIGESTIVE ORGANS AND PERITONEUM 5.878498
(code, 179-189) MALIGNANT NEOPLASM OF GENITOURINARY ORGANS 4.572165
(code, 540-543) APPENDICITIS 4.572165
(code, E870) Accidental cut, puncture, perforation, or hemorrhage during medical care 4.572165
(code, E936) Anticonvulsants and anti-Parkinsonism drugs 4.572165
(code, 160-165) MALIGNANT NEOPLASM OF RESPIRATORY AND INTRATHORACIC ORGANS 4.267354
(code, 782) Symptoms involving skin and other integumentary tissue 3.657732
(code, 320-326) INFLAMMATORY DISEASES OF THE CENTRAL NERVOUS SYSTEM 3.657732
(code, 781) Symptoms involving nervous and musculoskeletal systems 3.657732
(code, 050-059) VIRAL DISEASES ACCOMPANIED BY EXANTHEM 3.483554
(code, V10) Personal history of malignant neoplasm 3.457558
(code, V13) Personal history of other diseases 3.325211
Advanced.Heart.Disease
descr OR
icd9
(code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted 7.153226
(code, E941) Drugs primarily affecting the autonomic nervous system 7.153226
(code, 137-139) LATE EFFECTS OF INFECTIOUS AND PARASITIC DISEASES 7.153226
(code, 799.1) Respiratory arrest 7.153226
(code, E937) Sedatives and hypnotics 7.153226
(code, 958-959) CERTAIN TRAUMATIC COMPLICATIONS AND UNSPECIFIED INJURIES 4.768817
(code, E858) Accidental poisoning by other drugs 4.768817
(code, 791) Nonspecific findings on examination of urine 3.576613
(code, V53) Fitting and adjustment of other device 3.065668
(code, E942) Agents primarily affecting the cardiovascular system 2.861290
(code, E931) Other anti-infectives 2.384409
(code, 393-398) CHRONIC RHEUMATIC HEART DISEASE 2.322476
(code, V70) General medical examination 2.043779
(code, 410-414) ISCHEMIC HEART DISEASE 1.792237
(code, 796) Other nonspecific abnormal findings 1.788306
(code, E870) Accidental cut, puncture, perforation, or hemorrhage during medical care 1.788306
(code, 460-466) ACUTE RESPIRATORY INFECTIONS 1.788306
(code, 170-176) MALIGNANT NEOPLASM OF BONE, CONNECTIVE TISSUE, SKIN, AND BREAST 1.788306
(code, 600-608) DISEASES OF MALE GENITAL ORGANS 1.744689
(code, V45) Other postprocedural states 1.722073
Advanced.Lung.Disease
descr OR
icd9
(code, 980-989) TOXIC EFFECTS OF SUBSTANCES CHIEFLY NONMEDICINAL AS TO SOURCE 11.826667
(code, 791) Nonspecific findings on examination of urine 11.826667
(code, E869) Accidental poisoning by other gases and vapors 11.826667
(code, 799.1) Respiratory arrest 11.826667
(code, V46) Other dependence on machines and devices 7.446420
(code, 460-466) ACUTE RESPIRATORY INFECTIONS 5.174167
(code, V02) Carrier or suspected carrier of infectious diseases 4.435000
(code, V13) Personal history of other diseases 4.300606
(code, 730-739) OSTEOPATHIES, CHONDROPATHIES, AND ACQUIRED MUSCULOSKELETAL DEFORMITIES 4.108211
(code, 617-629) OTHER DISORDERS OF FEMALE GENITAL TRACT 3.942222
(code, E853) Accidental poisoning by tranquilizers 3.942222
(code, V14) Personal history of allergy to medicinal agents 3.942222
(code, E917) Striking against or struck accidentally by objects or persons 3.942222
(code, 958-959) CERTAIN TRAUMATIC COMPLICATIONS AND UNSPECIFIED INJURIES 3.942222
(code, 338) PAIN 3.379048
(code, V16) Family history of malignant neoplasm 3.379048
(code, E936) Anticonvulsants and anti-Parkinsonism drugs 2.956667
(code, 317-319) MENTAL RETARDATION 2.571014
(code, 490-496) CHRONIC OBSTRUCTIVE PULMONARY DISEASE AND ALLIED CONDITIONS 2.569241
(code, 249-259) DISEASES OF OTHER ENDOCRINE GLANDS 2.489825
Alcohol.Abuse
descr OR
icd9
(code, 840-848) SPRAINS AND STRAINS OF JOINTS AND ADJACENT MUSCLES 10.950617
(code, V62) Other psychosocial circumstances 10.950617
(code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted 10.950617
(code, E967) Perpetrator of child and adult abuse 10.950617
(code, V69) Problems related to lifestyle 10.950617
(code, V65) Other persons seeking consultation 10.950617
(code, E960) Fight, brawl, rape 10.950617
(code, E901) Excessive cold 10.950617
(code, V60) Housing, household, and economic circumstances 9.855556
(code, V11) Personal history of mental disorder 8.212963
(code, E858) Accidental poisoning by other drugs 7.300412
(code, V63) Unavailability of other medical facilities for care 5.475309
(code, 830-839) DISLOCATION 5.475309
(code, E854) Accidental poisoning by other psychotropic agents 5.475309
(code, E928) Other and unspecified environmental and accidental causes 4.211776
(code, 795) Other and nonspecific abnormal cytological, histological, immunological and DNA test findings 3.650206
(code, E931) Other anti-infectives 3.650206
(code, E853) Accidental poisoning by tranquilizers 3.650206
(code, 070-079) OTHER DISEASES DUE TO VIRUSES AND CHLAMYDIAE 3.629699
(code, V08) Asymptomatic human immunodeficiency virus [HIV] infection status 3.422068
Chronic.Neurological.Dystrophies
descr OR
icd9
(code, V88) ACQUIRED ABSENCE OF OTHER ORGANS AND TISSUE 6.694340
(code, 137-139) LATE EFFECTS OF INFECTIOUS AND PARASITIC DISEASES 6.694340
(code, E901) Excessive cold 6.694340
(code, 047) Meningitis due to enterovirus 6.694340
(code, E945) Agents primarily acting on the smooth and skeletal muscles and respiratory system 6.694340
(code, E936) Anticonvulsants and anti-Parkinsonism drugs 5.020755
(code, 617-629) OTHER DISORDERS OF FEMALE GENITAL TRACT 4.462893
(code, V62) Other psychosocial circumstances 3.347170
(code, V11) Personal history of mental disorder 3.347170
(code, 791) Nonspecific findings on examination of urine 3.347170
(code, 794) Nonspecific abnormal results of function studies 3.347170
(code, E929) LATE EFFECTS OF ACCIDENTAL INJURY 3.347170
(code, E935) Analgesics, antipyretics, and antirheumatics 3.347170
(code, 340-349) OTHER DISORDERS OF THE CENTRAL NERVOUS SYSTEM 3.047829
(code, V16) Family history of malignant neoplasm 2.869003
(code, 905-909) LATE EFFECTS OF INJURIES, POISONINGS, TOXIC EFFECTS, AND OTHER EXTERNAL CAUSES 2.869003
(code, V49) Other conditions influencing health status 2.818669
(code, 784) Symptoms involving head and neck 2.454591
(code, V13) Personal history of other diseases 2.434305
(code, 330-337) HEREDITARY AND DEGENERATIVE DISEASES OF THE CENTRAL NERVOUS SYSTEM 2.373448
Chronic.Pain.Fibromyalgia
descr OR
icd9
(code, V26) Procreative management 9.386243
(code, 338) PAIN 5.363568
(code, 791) Nonspecific findings on examination of urine 4.693122
(code, 614-616) INFLAMMATORY DISEASE OF FEMALE PELVIC ORGANS 4.693122
(code, E935) Analgesics, antipyretics, and antirheumatics 4.693122
(code, 170-176) MALIGNANT NEOPLASM OF BONE, CONNECTIVE TISSUE, SKIN, AND BREAST 3.519841
(code, 725-729) RHEUMATISM, EXCLUDING THE BACK 3.481994
(code, V13) Personal history of other diseases 3.413179
(code, 782) Symptoms involving skin and other integumentary tissue 3.128748
(code, E929) LATE EFFECTS OF ACCIDENTAL INJURY 3.128748
(code, 794) Nonspecific abnormal results of function studies 3.128748
(code, E917) Striking against or struck accidentally by objects or persons 3.128748
(code, E960) Fight, brawl, rape 3.128748
(code, 520-529) DISEASES OF ORAL CAVITY, SALIVARY GLANDS, AND JAWS 3.128748
(code, 730-739) OSTEOPATHIES, CHONDROPATHIES, AND ACQUIRED MUSCULOSKELETAL DEFORMITIES 2.865274
(code, V46) Other dependence on machines and devices 2.781109
(code, 905-909) LATE EFFECTS OF INJURIES, POISONINGS, TOXIC EFFECTS, AND OTHER EXTERNAL CAUSES 2.681784
(code, 710-719) ARTHROPATHIES AND RELATED DISORDERS 2.573647
(code, 720-724) DORSOPATHIES 2.479385
(code, V63) Unavailability of other medical facilities for care 2.346561
Dementia
descr OR
icd9
(code, E869) Accidental poisoning by other gases and vapors 26.088235
(code, 980-989) TOXIC EFFECTS OF SUBSTANCES CHIEFLY NONMEDICINAL AS TO SOURCE 26.088235
(code, V11) Personal history of mental disorder 6.522059
(code, 290-299) PSYCHOSES 5.748255
(code, 330-337) HEREDITARY AND DEGENERATIVE DISEASES OF THE CENTRAL NERVOUS SYSTEM 3.794652
(code, 930-939) EFFECTS OF FOREIGN BODY ENTERING THROUGH ORIFICE 3.726891
(code, 860-869) INTERNAL INJURY OF THORAX, ABDOMEN, AND PELVIS 3.726891
(code, V53) Fitting and adjustment of other device 3.726891
(code, E939) Psychotropic agents 2.898693
(code, 617-629) OTHER DISORDERS OF FEMALE GENITAL TRACT 2.898693
(code, E935) Analgesics, antipyretics, and antirheumatics 2.608824
(code, 781) Symptoms involving nervous and musculoskeletal systems 2.608824
(code, 380-389) DISEASES OF THE EAR AND MASTOID PROCESS 2.608824
(code, 600-608) DISEASES OF MALE GENITAL ORGANS 2.545194
(code, E888) Other and unspecified fall 2.484594
(code, 783) Symptoms concerning nutrition, metabolism, and development 2.371658
(code, 787) Symptoms involving digestive system 2.248986
(code, 790) Nonspecific findings on examination of blood 2.174020
(code, 500-508) PNEUMOCONIOSES AND OTHER LUNG DISEASES DUE TO EXTERNAL AGENTS 2.073237
(code, E928) Other and unspecified environmental and accidental causes 2.006787
Depression
descr OR
icd9
(code, E931) Other anti-infectives 5.704180
(code, E960) Fight, brawl, rape 5.704180
(code, V65) Other persons seeking consultation 5.704180
(code, E901) Excessive cold 5.704180
(code, E950.0) Analgesics, antipyretics, and antirheumatics 5.704180
(code, E953.8) Other specified means 5.704180
(code, V69) Problems related to lifestyle 5.704180
(code, E967) Perpetrator of child and adult abuse 5.704180
(code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted 5.704180
(code, 540-543) APPENDICITIS 4.278135
(code, V62) Other psychosocial circumstances 4.278135
(code, V60) Housing, household, and economic circumstances 3.707717
(code, E850) Accidental poisoning by analgesics, antipyretics, and antirheumatics 3.259531
(code, E854) Accidental poisoning by other psychotropic agents 2.852090
(code, 614-616) INFLAMMATORY DISEASE OF FEMALE PELVIC ORGANS 2.852090
(code, E942) Agents primarily affecting the cardiovascular system 2.852090
(code, 338) PAIN 2.716276
(code, V13) Personal history of other diseases 2.592809
(code, E939) Psychotropic agents 2.535191
(code, 960-979) POISONING BY DRUGS, MEDICINAL AND BIOLOGICAL SUBSTANCES 2.444649
Developmental.Delay.Retardation
descr OR
icd9
(code, 317-319) MENTAL RETARDATION 55.093168
(code, E939) Psychotropic agents 9.386243
(code, 327) ORGANIC SLEEP DISORDERS 6.812596
(code, 787) Symptoms involving digestive system 5.825944
(code, 320-326) INFLAMMATORY DISEASES OF THE CENTRAL NERVOUS SYSTEM 5.631746
(code, 690-698) OTHER INFLAMMATORY CONDITIONS OF SKIN AND SUBCUTANEOUS TISSUE 5.450077
(code, 500-508) PNEUMOCONIOSES AND OTHER LUNG DISEASES DUE TO EXTERNAL AGENTS 3.356670
(code, 240-246) DISORDERS OF THYROID GLAND 3.320025
(code, 555-558) NONINFECTIOUS ENTERITIS AND COLITIS 3.312792
(code, 150-159) MALIGNANT NEOPLASM OF DIGESTIVE ORGANS AND PERITONEUM 3.017007
(code, 780) General symptoms 2.829346
(code, 340-349) OTHER DISORDERS OF THE CENTRAL NERVOUS SYSTEM 2.747193
(code, V55) Attention to artificial openings 2.639881
(code, 001-009) INTESTINAL INFECTIOUS DISEASES 2.509194
(code, 550-553) HERNIA OF ABDOMINAL CAVITY 2.283140
(code, 393-398) CHRONIC RHEUMATIC HEART DISEASE 2.194187
(code, 990-995) OTHER AND UNSPECIFIED EFFECTS OF EXTERNAL CAUSES 2.035571
(code, 590-599) OTHER DISEASES OF URINARY SYSTEM 2.011338
(code, 700-709) OTHER DISEASES OF SKIN AND SUBCUTANEOUS TISSUE 1.928680
(code, 249-259) DISEASES OF OTHER ENDOCRINE GLANDS 1.905478
Non.Adherence
descr OR
icd9
(code, E901) Excessive cold 13.238806
(code, 047) Meningitis due to enterovirus 13.238806
(code, E967) Perpetrator of child and adult abuse 13.238806
(code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted 13.238806
(code, E950.0) Analgesics, antipyretics, and antirheumatics 13.238806
(code, E960) Fight, brawl, rape 8.825871
(code, E858) Accidental poisoning by other drugs 8.825871
(code, E854) Accidental poisoning by other psychotropic agents 6.619403
(code, 614-616) INFLAMMATORY DISEASE OF FEMALE PELVIC ORGANS 6.619403
(code, V62) Other psychosocial circumstances 6.619403
(code, 960-979) POISONING BY DRUGS, MEDICINAL AND BIOLOGICAL SUBSTANCES 4.728145
(code, 794) Nonspecific abnormal results of function studies 4.412935
(code, V85) BODY MASS INDEX 4.412935
(code, E931) Other anti-infectives 4.412935
(code, V60) Housing, household, and economic circumstances 3.971642
(code, V15) Other personal history presenting hazards to health 3.501172
(code, 540-543) APPENDICITIS 3.309701
(code, V18) Family history of certain other specific conditions 3.309701
(code, 338) PAIN 3.152097
(code, E939) Psychotropic agents 2.941957
None
descr OR
icd9
(code, V88) ACQUIRED ABSENCE OF OTHER ORGANS AND TISSUE 1.650233
(code, V50) Elective surgery for purposes other than remedying health states 1.650233
(code, 793) Nonspecific abnormal findings on radiological and other examination of body structure 1.650233
(code, 900-904) INJURY TO BLOOD VESSELS 1.650233
(code, E917) Striking against or struck accidentally by objects or persons 1.650233
(code, E938) Other central nervous system depressants and anesthetics 1.650233
(code, E940) Central nervous system stimulants 1.650233
(code, E944) Water, mineral, and uric acid metabolism drugs 1.650233
(code, E950.2) Other sedatives and hypnotics 1.650233
(code, E950.3) Tranquilizers and other psychotropic agents 1.650233
(code, E957.1) Other man-made structures 1.650233
(code, V07) Need for isolation and other prophylactic measures 1.650233
(code, V14) Personal history of allergy to medicinal agents 1.650233
(code, 239) NEOPLASMS OF UNSPECIFIED NATURE 1.650233
(code, E945) Agents primarily acting on the smooth and skeletal muscles and respiratory system 1.650233
(code, 120-129) HELMINTHIASES 1.650233
(code, 062) Mosquito-borne viral encephalitis 1.650233
(code, V59) Donors 1.650233
(code, 137-139) LATE EFFECTS OF INFECTIOUS AND PARASITIC DISEASES 1.650233
(code, 920-924) CONTUSION WITH INTACT SKIN SURFACE 1.443953
Obesity
descr OR
icd9
(code, E950.0) Analgesics, antipyretics, and antirheumatics 22.175000
(code, V85) BODY MASS INDEX 7.391667
(code, E917) Striking against or struck accidentally by objects or persons 7.391667
(code, V18) Family history of certain other specific conditions 5.543750
(code, 327) ORGANIC SLEEP DISORDERS 3.934274
(code, 317-319) MENTAL RETARDATION 3.856522
(code, E929) LATE EFFECTS OF ACCIDENTAL INJURY 3.695833
(code, 720-724) DORSOPATHIES 3.347170
(code, V16) Family history of malignant neoplasm 3.167857
(code, 905-909) LATE EFFECTS OF INJURIES, POISONINGS, TOXIC EFFECTS, AND OTHER EXTERNAL CAUSES 3.167857
(code, 130-136) OTHER INFECTIOUS AND PARASITIC DISEASES 3.023864
(code, 320-326) INFLAMMATORY DISEASES OF THE CENTRAL NERVOUS SYSTEM 2.956667
(code, 920-924) CONTUSION WITH INTACT SKIN SURFACE 2.771875
(code, 415-417) DISEASES OF PULMONARY CIRCULATION 2.589781
(code, 700-709) OTHER DISEASES OF SKIN AND SUBCUTANEOUS TISSUE 2.531393
(code, 680-686) INFECTIONS OF SKIN AND SUBCUTANEOUS TISSUE 2.491573
(code, 617-629) OTHER DISORDERS OF FEMALE GENITAL TRACT 2.463889
(code, E942) Agents primarily affecting the cardiovascular system 2.217500
(code, 781) Symptoms involving nervous and musculoskeletal systems 2.217500
(code, 710-719) ARTHROPATHIES AND RELATED DISORDERS 2.145968
Other.Substance.Abuse
descr OR
icd9
(code, E960) Fight, brawl, rape 16.735849
(code, 840-848) SPRAINS AND STRAINS OF JOINTS AND ADJACENT MUSCLES 16.735849
(code, 062) Mosquito-borne viral encephalitis 16.735849
(code, E950.0) Analgesics, antipyretics, and antirheumatics 16.735849
(code, E980) Poisoning by solid or liquid substances, undetermined whether accidentally or purposely inflicted 16.735849
(code, E854) Accidental poisoning by other psychotropic agents 16.735849
(code, V69) Problems related to lifestyle 16.735849
(code, E967) Perpetrator of child and adult abuse 16.735849
(code, V60) Housing, household, and economic circumstances 11.715094
(code, E850) Accidental poisoning by analgesics, antipyretics, and antirheumatics 9.563342
(code, 960-979) POISONING BY DRUGS, MEDICINAL AND BIOLOGICAL SUBSTANCES 9.563342
(code, 070-079) OTHER DISEASES DUE TO VIRUSES AND CHLAMYDIAE 5.923362
(code, E939) Psychotropic agents 5.578616
(code, E853) Accidental poisoning by tranquilizers 5.578616
(code, E858) Accidental poisoning by other drugs 5.578616
(code, 795) Other and nonspecific abnormal cytological, histological, immunological and DNA test findings 5.578616
(code, E931) Other anti-infectives 5.578616
(code, E928) Other and unspecified environmental and accidental causes 5.149492
(code, 042) Human immunodeficiency virus [HIV] disease 4.462893
(code, V62) Other psychosocial circumstances 4.183962
Schizophrenia.and.other.Psychiatric.Disorders
descr OR
icd9
(code, E901) Excessive cold 8.959596
(code, V69) Problems related to lifestyle 8.959596
(code, E958.9) Unspecified means 8.959596
(code, E960) Fight, brawl, rape 5.973064
(code, V85) BODY MASS INDEX 5.973064
(code, E853) Accidental poisoning by tranquilizers 5.973064
(code, V60) Housing, household, and economic circumstances 5.823737
(code, E939) Psychotropic agents 4.977553
(code, 540-543) APPENDICITIS 4.479798
(code, 850-854) INTRACRANIAL INJURY, EXCLUDING THOSE WITH SKULL FRACTURE 4.479798
(code, V13) Personal history of other diseases 4.072544
(code, E850) Accidental poisoning by analgesics, antipyretics, and antirheumatics 3.839827
(code, 338) PAIN 3.839827
(code, 317-319) MENTAL RETARDATION 3.116381
(code, 958-959) CERTAIN TRAUMATIC COMPLICATIONS AND UNSPECIFIED INJURIES 2.986532
(code, E917) Striking against or struck accidentally by objects or persons 2.986532
(code, 795) Other and nonspecific abnormal cytological, histological, immunological and DNA test findings 2.986532
(code, E858) Accidental poisoning by other drugs 2.986532
(code, 290-299) PSYCHOSES 2.935913
(code, 470-478) OTHER DISEASES OF THE UPPER RESPIRATORY TRACT 2.780564
Unsure
descr OR
icd9
(code, E945) Agents primarily acting on the smooth and skeletal muscles and respiratory system 5.142029
(code, E937) Sedatives and hypnotics 5.142029
(code, 137-139) LATE EFFECTS OF INFECTIOUS AND PARASITIC DISEASES 5.142029
(code, E858) Accidental poisoning by other drugs 3.428019
(code, V16) Family history of malignant neoplasm 2.938302
(code, 617-629) OTHER DISORDERS OF FEMALE GENITAL TRACT 2.856683
(code, 540-543) APPENDICITIS 2.571014
(code, E936) Anticonvulsants and anti-Parkinsonism drugs 2.571014
(code, 791) Nonspecific findings on examination of urine 2.571014
(code, 830-839) DISLOCATION 2.571014
(code, 320-326) INFLAMMATORY DISEASES OF THE CENTRAL NERVOUS SYSTEM 2.399614
(code, E939) Psychotropic agents 2.285346
(code, V46) Other dependence on machines and devices 2.094901
(code, E934) Agents primarily affecting blood constituents 2.012098
(code, V08) Asymptomatic human immunodeficiency virus [HIV] infection status 1.928261
(code, 725-729) RHEUMATISM, EXCLUDING THE BACK 1.824591
(code, 730-739) OSTEOPATHIES, CHONDROPATHIES, AND ACQUIRED MUSCULOSKELETAL DEFORMITIES 1.786178
(code, E931) Other anti-infectives 1.714010
(code, V14) Personal history of allergy to medicinal agents 1.714010
(code, 520-529) DISEASES OF ORAL CAVITY, SALIVARY GLANDS, AND JAWS 1.714010


Comments

comments powered by Disqus