Essential Hypertension

In addition to identifying a cause list and mapping this cause list across various revisions of the ICD, the largest impediment to comparability is the presence of a different set of GCs in each ICD revision. To more fully understand the problem of garbage codes, we created a typology of these codes that distinguishes four types of GCs. This typology has been developed taking into consideration the following: the likelihood that a condition can be an underlying cause of death; the need for codes that provide a location for unspecified or ambiguous causes of death; and the need for codes that represent causes that are not underlying but intermediate or final events in the chain leading to death. Four categories were identified:
1. Causes that cannot or should not be considered as underlying causes of death. These are codes that are included in the ICD because of its use for classifying health service encounters but that do not signify underlying cause of death. Examples of this type of GC are all the codes under chapter 18 of ICD-10 or R codes. This category also includes two special cases in the cardiovascular area: essential primary hypertension and atherosclerosis. Essential primary hypertension is included in the ICD to classify clinical encounters, but for most physicians, it should be considered a risk factor for cardiovascular disease and not the underlying cause. This distinction between what is a risk factor and what is an underlying cause is somewhat arbitrary but necessary to enhance comparability across revisions. Finally, we included in this category a number of causes that are described as the long-term sequelae of disease, such as G82, paraplegia and tetraplegia, or O94, sequelae of complication of pregnancy, childbirth, and the puerperium. In these cases, for public health purposes, it is more useful to assign these deaths to the underlying cause despite the long time lag between disease and death.
2. Intermediate causes of death such as heart failure, septicemia, peritonitis, osteomyelitis, or pulmonary embolism. These are clearly defined clinical entities, but each has an underlying cause that would have precipitated the chain of events leading to death. Physicians who have not been adequately trained in the principles of the ICD underlying cause of death often use these causes on death certificates.
3. Immediate causes of death that are the final steps in a disease pathway leading to death. Examples of this include disseminated intravascular coagulation or defibrination syndrome (D65). The pathway to death includes the final immediate cause, an intermediate cause, and the underlying cause that triggered the chain of events. Cardiac arrest (I46) and respiratory failure, not elsewhere classified (J96), are other examples.
4. Unspecified causes within a larger cause grouping. For many diseases, such as neoplasms, a code is included within the grouping for an unspecified site. This is an illustration of a GC that is not important for assessing aggregate deaths from neoplasms from all sites but is important when assessing site-specific death rates. Another important example is the injury category in which some injuries are coded to unspecified factors or intent.
Table 2 provides a listing of the number of each type of GC that we identified related to our 56-cause list. The largest category of GCs is type 1. Assessment of the number of GCs, especially in category 4, is a function of the level of detail in the final cause list that is being developed.

We extracted ICD-10-CM codes from the 2018AA release of the UMLS [18 ] and used several automated methods to translate ICD-10-CM diagnosis codes to phecodes (Figure 1). We mapped 515 ICD-10-CM codes directly to phecodes by matching code descriptions regardless of capitalization (eg, ICD-10-CM H52.4 “Presbyopia” to phecode 367.4 “Presbyopia”). We mapped 82,287 ICD-10-CM codes indirectly to phecodes using the existing ICD-9-CM phecode map [14 ]. To convert ICD-10-CM codes indirectly to phecodes, we used General Equivalence Mappings (GEMS) provided by the Centers for Medicare & Medicaid Services that map ICD-10-CM to ICD-9-CM and vice versa [22 ]. We included both equivalent and nonequivalent GEMS mappings (ie, where the approximate flag was either 0 or 1). As an example of this indirect approach, to map ICD-10-CM E11.9 “Type 2 diabetes mellitus without complications” to phecode 250.2 “Type 2 diabetes,” we mapped ICD-10-CM E11.9 to ICD-9-CM 250.0 “Diabetes mellitus without mention of complication” to phecode 250.2.
Since the GEMS do not provide ICD-9-CM mappings for all ICD-10-CM codes [17 (link)], we complemented this approach with UMLS semantic mapping [24 (link)], Observational Health Data Sciences and Informatics (OHDSI) concept relationships [25 ,26 (link)], and National Library of Medicine (NLM) maps [23 ]. In this approach to indirect mapping, we first mapped ICD-10-CM codes to Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) through UMLS Concept Unique Identifier (CUI) equivalents, which were then converted to ICD-9-CM through either UMLS CUI equivalents [18 ,24 (link)], OHDSI [25 ], or NLM maps [23 ]. For example, we mapped ICD-10-CM L01.00 “Impetigo, unspecified” to CUI C0021099 to SNOMED CT 48277006 to OHDSI Concept ID 140480 to OHDSI Concept ID 44832600 to ICD-9-CM 684 and finally to phecode 686.2 “Impetigo”.
There were two general instances when an ICD-10-CM code mapped to more than one phecode. First, some ICD-10-CM codes mapped to both a parent phecode and one of its child phecodes that was lower in the hierarchy. To maintain the granular meanings of ICD-10-CM codes, we only kept the mappings to child phecodes, a decision that we could make due to the hierarchical structure of phecodes. For example, ICD-10-CM I10 “Essential (primary) hypertension” was mapped to phecodes 401 “Hypertension” and 401.1 “Essential hypertension”, but we only kept the mapping to phecode 401.1. Second, we kept all the mappings for ICD-10-CM codes that were translated to phecodes that were not in the same family. This can be seen in the mapping of ICD-10-CM D57.812 “Other sickle-cell disorders with splenic sequestration” to phecodes 282.5 “Sickle cell anemia” and 289.5 “Diseases of spleen”. This latter association created a polyhierarchical nature to phecodes that did not previously exist.
To map ICD-10 (non-CM) codes to phecodes, we also used ICD-10 codes from the 2018AA UMLS [18 ]. ICD-10 codes were mapped to phecodes in a similar manner to ICD-10-CM, but since a GEMS to translate ICD-10 to ICD-9-CM was not available, we used only string matching and previously manually reviewed resources from the UMLS [24 (link)], NLM [23 ], and OHDSI [25 ,26 (link)].

Preoperative factors were collected, including age, sex, recent major cardiovascular procedure (within 3 months), coronary artery disease, cerebral vascular events, chronic lung disease, essential hypertension, dyslipidemia, liver cirrhosis, atrial fibrillation, type 2 diabetes mellitus, end-stage renal disease with dialysis (both hemodialysis and peritoneal dialysis), and regular use of antiplatelet or anticoagulant agents. The major cardiovascular procedures included coronary arterial bypass, coronary arterial angioplasty/stenting, cardiac valvular surgery, aortic surgery, and peripheral arterial surgery. Preoperative blood cell counts included white cell counts, differential counts (immature band form white cell) [15 (link)], platelet counts, the neutrophil-to-lymphocyte ratio (NLR) [16 (link)], and hemoglobin levels. Preoperative blood biochemistry results included serum levels of albumin, alanine aminotransferase (ALT), bilirubin, and creatinine. The coagulation test included the prothrombin time (PT) and was expressed by the international normalized ratio (INR). Preoperative shock status was defined as the requirement for vasopressors or inotropes. The types of AMI were determined by preoperative contrast CT scans.