The primary data source for the information on this website is the California inpatient discharge database. This data is made available by the California Department of Health Care Access and Information (HCAI) formerly the California Office of Statewide Planning and Development (OSHPD). The inpatient discharge database is based on data collected from licensed acute care California hospitals. A single record in the inpatient discharge data corresponds to a single inpatient encounter in a California hospital and includes clinical, patient demographic, payer and disposition information for this encounter.

The California inpatient discharge data base published by HCAI does not include discharges from military facilities or unlicensed facilities.

The annual California inpatient discharge data includes information on almost 3 million encounters of patients 18 or older in 443 acute care hospitals.

The California inpatient discharge database is used for a number of reports on health care delivery volume, health care quality and challenges in California: HCAI Health Care Quality reports.

For the identification of PCIs that were performed in an outpatient and emergency department setting, in addition to inpatient discharge data, data collected from hospital emergency departments and licensed ambulatory surgery clinics in California is used.

The California inpatient discharge data, emergency department and ambulatory surgery data offer a number of advantages:

• Reporting of inpatient and outpatient data is mandatory for all California licensed acute care hospitals (https://hcai.ca.gov/about/laws-regulations/).
• The majority of California inpatient discharges are included.
• Demographic patient information such as age, gender, race, ethnicity, primary language spoken are included.
• A principal diagnosis and up to 24 secondary diagnoses are reported. Diagnoses for discharges through September 30^th 2015, are reported using the ICD-9-CM coding system; diagnoses for discharges starting from October 1, 2015 are reported using the ICD-10-CM coding system; each diagnosis is reported with a "condition present at admission" (CPOA) indicator that allows the user to distinguish whether a condition was present at the time of hospital admission or developed during the admission.
• For inpatient data, a principal procedure and up to 20 - starting with 2019, up to 24 - secondary procedures are reported. Procedure for discharges through September 30^th 2015, are reported using the ICD-9-PCS coding system; procedures for discharges starting from October 1, 2015 are reported using the ICD-10-PCS coding system; for each procedure, the database also includes information on the timing of the procedure relative to the admission date.
• For outpatient data, a principal procedure and up to 20 - starting with 2019, up to 24 - secondary procedures are reported according to the Current Procedural Terminology (CPT-4) coding system.
• For inpatient data, admission source, admission type (non-elective vs. elective), and disposition information is available.
• For outpatient data, disposition information is available.
• Patients with multiple inpatient, emergency department and outpatient encounters can be connected over time.

Using the California inpatient discharge data, emergency department data and ambulatory surgery data does come with some disadvantages:

• The California inpatient discharge and outpatient data are administrative data bases primarily designed to support hospital billings rather than the study of health care outcomes. For this reason, there might be an incentive by hospitals to only include diagnostic and procedure codes in inpatient discharge and outpatient records that affect the hospital bill. Conditions with little or no impact on charges might not be reliably coded.
• While the California inpatient and outpatient data are rigorously checked for consistency, there is no regular data audit that evaluates coding quality.
• The California inpatient and outpatient data bases do not include clinical measurements (e.g., lab values), but solely relies on coding of diagnoses and procedures as ICD-9-CM, ICD-9-PCS, ICD-10-CM, ICD-10-PCS or CPT-4 codes.
• The California inpatient and outpatient data bases are limited to 25 diagnosies and 21 procedures for 2018 and earlier / 25 procedures for 2019 and later which might not be sufficient to reflect severely ill patients.
• The addition of "condition present at admission" (CPOA) indicators to the inpatient discharge data greatly increased the ability to assess the incidence of adverse events after hospital admission such as post-operative stroke. The accurate reporting of the CPOA indicator for each diagnosis in the inpatient discharge record is important for reporting of adverse events during the hospital stay. In a memorandum report Assessment of Hospital Reporting of Present at Admission Indicators on Medicare claims found that in a sample of 698 claims with 5,491 POA indicators 3% of the coded indicators were incorrect.
• As military hospitals or other unlicensed health facilities do not report to HCAI, any procedures performed at these institutions are not included in this analysis.

Please also review the section Why do the PCI volumes reported for a facility differ from volumes reported by other data sources? as we discovered possible under-reporting of outpatient PCIs by some facilities.

Data for 1999 through 2021 are included in this study. All statistical models were based on the most recent 2 years, '.$maxyear_1.' and '.$maxyear.'. Note that all year references pertain to the year of discharge (not the year of surgery).

To identify Coronary Artery Bypass Graft Surgeries, for discharges through September 30, 2015, each inpatient record was queried for the following ICD-9-PCS surgery codes: 36.11, 36.12, 36.13, 36.14, 36.15, 36.16, 36.17, 36.19.

For discharges starting from October 1, 2015, each inpatient record was queried for the following ICD-10-PCS surgery codes: 0210xxx, 0211xxx, 0212xxx, 0213xxx. For instance, all procedure codes starting with 0210 were included to identify CABG surgeries irrespective of characters in positions 5 through 7 of the ICD-10-PCS code.

To identify isolated CABG surgeries, the approach adopted by the California Coronary Artery Bypass Graft (CABG) Outcomes Reporting Program (CCORP) was followed:

"Isolated CABG surgeryis defined as CABG surgery performed on patients aged 18 years or older without other major procedures, such as valve repair or carotid endarterectomy, during the same surgery."
see 2016 Technical Note, page 2

Per the CCORP Training Manual (pages 71-73), the co-occurence of the following procedures on the day of the CABG make the CABG procedure non-isolated:

• Valve repairs or replacements
• Operations on structures adjacent to heart valves (papillary muscle, chordae tendineae, traebeculae carneae cordis, annuloplasty, infundibulectomy)
• Ventriculectomy when diagnosed preoperatively as a rupture, aneurysm or remodeling procedure. But not 1) sites intra-operatively diagnosed, 2) patch applications for site oozing discovered during surgery and 3) prophylactic patch applications to reduce chances of future rupture
• Repair of atrial and ventricular septa, but not closure of patent foramen ovale
• Excision of aneurysm of heart
• Head and neck, intracranial endarterectomy
• Other open heart surgeries, such as aortic arch repair, pulmonary endarterectomy
• Endarterectomy of aorta
• Thoracic endarterectomy (endarterectomy on an artery outside the heart)
• Carotid endarterectomy
• Heart transplantation
• Repair of certain congenital cardiac anomalies, but not closure of patent foramen ovale (e.g., teratology of fallot, atrial septal defect (ASD), ventricular septal defect (VSD), valvular abnormality)
• Any aortic aneurysm repair (abdominal or thoracic)
• Aorta-subclavian-carotid bypass
• Aorta-renal bypass
• Aorta-iliac-femoral bypass
• Caval-pulmonary artery anastomosis
• Extracranial-intracranial (EC-IC) vascular bypass
• [Repair of] coronary artery fistula
• Resection of a lobe or segment of the lung (e.g., lobectomy or segmental resection of lung). But not simple biopsy of lung nodule in which surrounding lung is not resected, biopsy of a thoracic lymph node or excision or stapling of an emphysematous bleb.
• Pleural decortication
• Mastectomy for breast cancer (not simple breast biopsy)
• Amputation of any extremity (e.g., foot or toe)
• Resection of LV aneurysm
• Ventricular Assist Device (VAD) as bridge to transplant
• Septal myectomy with hypertrophic obstructive cardiomyopathy
• Full open mazes
• Repair of aortic dissection

The identification of the presence of exclusionary conditions has to be accomplished based on the procedure and possibly diagnosis codes provided in a patient's inpatient discharge record which poses many challenges. Click here to review ICD-9-PCS and ICD-10-PCS codes used to assess isolated vs. non-isolated status.

For discharges through September 30, 2015, each inpatient record was queried for the following ICD-9-PCS surgery codes:

Surgical Aortic Valve Replacement	35.21, 35.22
Surgical Mitral Valve Replacement	35.23, 35.24
Surgical Mitral Valve Repair	35.12, 35.33
Transcatheter Aortic Valve Replacement (TAVR)	35.05, 35.06
Transcatheter Mitral Valve Replacement	No codes
Transcatheter Mitral Valve Repair	35.96, 35.97

For discharges starting from October 1, 2015, each inpatient record was queried for the following ICD-10-PCS surgery codes:

Surgical Aortic Valve Replacement	02RF0xx, X2RF0xx and ECC-code 5A1221Z on the date of surgery
Surgical Mitral Valve Replacement	02RG0xx and ECC-code 5A1221Z on the date of surgery
Surgical Mitral Valve Repair	027G0xx, 02QG0xx, 02UG0xx, 02NG0xx, 02VG0xx, 02WG0xx and ECC-code 5A1221Z on the date of surgery
Transcatheter Aortic Valve Replacement	02RF3xx and absence of ECC-code 5A1221Z on the date of surgery and absence of CABG surgery code on the date of surgery
Transcatheter Mitral Valve Replacement	02RG3xx and absemce pf ECC-code 5A1221Z on the date of surgery
Transcatheter Mitral Valve Repair	027G3xx, 02QG3xx, 02UG3xx, 02NG3xx, 02VG3xx, 02WG3xx and absence of ECC-code 5A1221Z on the date of surgery

The group of patients with surgical AV-Replacement, MV-Replacement or MV-Repair was further divided according to whether a valve surgery and a CABG surgery were performed on the same day, resulting in the group of isolated valve surgeries and valve plus CABG surgeries.

As for the identification for non-isolated CABG surgeries, isolated procedures were identified by checking for the co-occurrence of the following procedures on the date of the valve or valve/CABG procedure (see CCORP Training Manual (pages 71-73)).

• Aortic Valve repair
• Aortic Valve root replacement with valved conduit (Bentall)
• Ventriculectomy when diagnosed preoperatively as a rupture, aneurysm or remodeling procedure. But not 1) sites intra-operatively diagnosed, 2) patch applications for site oozing discovered during surgery and 3) prophylactic patch applications to reduce chances of future rupture
• Repair of atrial and ventricular septa, but not closure of patent foramen ovale
• Excision of aneurysm of heart
• Head and neck, intracranial endarterectomy
• Other open heart surgeries, such as aortic arch repair, pulmonary endarterectomy
• Endarterectomy of aorta
• Thoracic endarterectomy (endarterectomy on an artery outside the heart)
• Carotid endarterectomy
• Heart transplantation
• Repair of congenital cardiac anomalies, such as tetralogy of fallot, atrial septal defect (ASD), ventricular septal defect or other complex anomaly
• Any aortic aneurysm repair (abdominal or thoracic)
• Repair of aortic dissection
• Aorta-subclavian-carotid bypass
• Aorta-renal bypass
• Aorta-iliac-femoral bypass
• Caval-pulmonary artery anastomosis
• Extracranial-intracranial (EC-IC) vascular bypass
• [Repair of] coronary artery fistula
• Resection of a lobe or segment of the lung (e.g., lobectomy or segmental resection of lung). But not simple biopsy of lung nodule in which surrounding lung is not resected, biopsy of a thoracic lymph node or excision or stapling of an emphysematous bleb.
• Pleural decortication
• Mastectomy for breast cancer (not simple breast biopsy)
• Amputation of any extremity (e.g., foot or toe)
• Resection of LV aneurysm
• Ventricular Assist Device (VAD) as a bridge to transplant
• Infundibulectomy
• Septal myectomy with hypertrophic obstructive cardiomyopathy
• Full Open MAZE for Aortic Valve cases only

Note that in contrast to CCORP, CCSIP is not considering the occurrence of a pulmonic or tricuspid valve procedure an exclusion.

The identification of the presence of exclusionary conditions has to be accomplished based on the procedure and possibly diagnosis codes provided in a patient's inpatient discharge record which poses many challenges. Click here to review ICD-9-PCS and ICD-10-PCS codes used to assess isolated vs. non-isolated status.

For discharges through September 30, 2015, each inpatient record was queried for the following ICD-9-PCS surgery codes: 36.01, 36.02, 36.05, 36.06, 36.07, 36.09, 00.66, 17.55.

For discharges starting from October 1, 2015, each inpatient record was queried for the following ICD-10-PCS surgery codes: 02703xx, 02713xx, 02723xx, 02733xx, 02C03xx, 02C13xx, 02C23xx, 02C33xx, X2C03xx, X2C13xx, X2C23xx, X2C33xx, 02704xx, 02714xx, 02724xx, 02734xx, 02C04xx, 02C14xx, 02C24xx, 02C34xx, X2C04xx, X2C14xx, X2C24xx, X2C34xx.

CPT coding for PCIs was changed in 2013. For 2005 through 2012, each ambulatory surgery and emergency department record was queried for the following CPT codes: 92980, 92981, 92982, 92984, 92995, 92996.

For 2013 and later, each ambulatory surgery and emergency department record was queried for the following CPT codes: 92920, 92921, 92924, 92925, 92928, 92929, 92933, 92934, 92937, 92938, 92941, 92943, 92944.

Within the groups of patients with a PCI, we furthermore distinguished patients with Acute Coronary Syndrome (ACS) and patients without ACS. For discharges through September 30, 2015, if an inpatient record included a diagnosis code for acute MI (410.0-410.8) coded as present at admission (POA=Yes) at the PCI admission, the case was considered PCI with ACS. For discharges starting from October 1, 2015, if an inpatient record included a diagnosis code for acute MI (I210x, I211x, I212x, I213, I220, I221, I228, I229, I214, I222) coded as present at admission (POA=Yes) at the PCI admission, the case was considered PCI with ACS. For an ambulatory surgery or emergency department record, a patient admitted with an acute MI diagnosis was placed in the PCI with ACS group.

For discharges through September 30, 2015, the following procedures are included:

For discharges starting from October 1, 2015, the following procedures are included:

The following table shows the volume of CABG and valve procedures that were excluded due to the presence of an exclusionary condition (see Sections How were isolated CABGs identified? and How were surgical and transcatheter aortic and mitral valve surgeries with or without CABG identfied?) for the most recent 2-year periods. The time periods were intentionally chosen to allow assessment of the impact of exclusions over time keeping in mind that diagnosis and procedure codes in the HCAI inaptient discharge data was changed from ICD-9 to ICD-10 starting with October 1, 2015. For all surgical groups studied, the percent of records excluded is increasing over time. Some of the increase might be due to the coding change. The substantial increase in the percent of records excluded for the Surgical AVR group might also be affected by the increasing use of TAVR such that patients with Surgical AVR have more complex disease issues more likely to include an exclusionary condition.

To get an idea of the volume of other surgical procedures, the table below presents the volume of other major cardiac procedures performed with ECC that are included in the Other Major Cardiac Surgery group and occurred with a frequency ≥ 50 in 2013-2014.

Volume of all Cardiac Procedures

Isolated CABG

Volume of isolated CABG procedures in the absence of any exclusionary condition as specified in the Section How were isolated CABGs identified?

Isolated Surgical or Transcatheter Aortic and Mitral Valve Procedures with or without CABG

Volume of Surgical and Transcatheter Aortic Valve Replacements, Mitral Valve Replacement, Mitral Valve Repair, multiple surgical valve procedures including AV-Replacement, MV-Replacement or MV-Repair with or without CABG in the absence of any exclusionary condition as specified in Section How were surgical and transcatheter aortic and mitral valve surgeries with or without CABG identfied? or Transcatheter AVR. It is possible that other valve procedures were done along with AV-Replacement, MV-Replacement or MV-Repair, however, in order to be included in this group, at least 1 of AV-Replacement, MV-Replacement or MV-Repair had to be performed.

PCI without ACS

Volume of percutaneous coronary interventions with no evidence of acute MI. See section How were Percutaneous Coronary Interventions (PCI) identified?

PCI with ACS

Volume of percutaneous coronary interventions with evidence of acute MI. See section How were Percutaneous Coronary Interventions (PCI) identified?

Other Cardiac Surgeries

Volume of all other major cardiac surgeries requiring ECC, pulmonic, tricuspid and other valve surgeries without AV-Replacement, MV-Replacement or MV-Repair, and non-isolated CABG and valve procedures with an exclusionary condition that make them ineligible for one of the previous groups. Also see Section What is included in Other Major Cardiac Procedures?.

Volume of Valve Procedures

Surgical AV-Replacement

Surgical aortic valve replacement in the absence of an exclusionary condition.

Surgical AV-Replacement + CABG

Surgical aortic valve replacement with CABG in the absence of an exclusionary condition.

Transcatheter AV-Replacement (TAVR)

Transcatheter aortic valve replacement in the absence of an exclusionary condition.

Surgical MV-Replacement

Surgical mitral valve replacement in the absence of an exclusionary condition.

Surgical MV-Replacement + CABG

Surgical mitral valve replacement with CABG in the absence of an exclusionary condition.

Transcatheter MV-Replacement

Transcatheter mitral valve replacement in the absence of an exclusionary condition.

Surgical MV-Repair

Surgical mitral valve repair in the absence of an exclusionary condition.

Surgical MV-Repair + CABG

Surgical mitral valve repair with CABG in the absence of an exclusionary condition.

Transcatheter MV-Repair

Transcatheter mitral valve repair in the absence of an exclusionary condition.

Multiple Surgical Valves

Multiple surgical valve procedures including at least two of AV-Replacement, MV-Replacement or MV-Repair in the absence of an exclusionary condition.

Multiple Surgical Valves + CABG

CABG plus multiple surgical valve procedures including at least two of AV-Replacement, MV-Replacement or MV-Repair in the absence of an exclusionary condition.

Other Surgical Valve

All other valve procedures or surgical AV-Replacement, MV-Replacement or MV-Repair with an exclusionary condition present.

Other Transcatheter Valve

All other transcatheter valve procedures, or transcatheter AV-Replacement, MV-Replacement or MV-Repair with an exclusionary condition present, or multiple transcatheter AV-Replacement, MV-Replacement or MV-Repair. Note that the number of multiple transcatheter valve procedures was very small and was therefore included in this group.

The HCAI inpatient discharge record includes the patient disposition. For patients transferred out, it is furthermore possible to find the subsequent patient admission based on a record linkage number that uniquely identifies the patient. We therefore included in our mortality outcome deaths that occurred after the index admission or death that occurred after a subsequent through-transfer-connected acute care admission. Patients who were discharged home, to a Skilled Nursing Facility (SNF), to long term care, to home health service or to a nursing home are considered survivors. Two admissions are considered connected if the disposition at the transferring facility is acute care or if the admission source at the facility a patient is transferred to is acute care and the type of care provided is acute inpatient care. Additionally, the discharge date at the transferring facility has to be within 1 day of the admission date at the facility a patient is transferred to.

The table below shows several ways of defining mortality after surgery and the implied mortality rates for California overall in 2010. Depending on the outcome chosen, mortality rates differ substantially. Of all measures - within the confines of data available and short of a primary data collection - measure (5) is the most complete and desirable. This measure includes mortality after the index admission, mortality after connected subsequent acute care admissions and deaths that occurred within 30 days of surgery irrespective of location. Unfortunately, this measure also hinges upon availability of the California Vital Statistics Death File which is typically available with a 3-year or longer gap making it unsuitable for timely outcome assessment.

CCSIP has adopted measure (3) as the mortality outcome measure:

Note that measure (5) in the table below comes close to the STS measure "operative mortality":

Using measure (1) substantially affects the mortality assessment especially in the group of patients undergoing PCI, a problem that can be addressed to a considerable extent by using measure (2) for the CABG, Surgical AV-Replacement, MV-Replacement, MV-Repair with CABG and Valve groups. For the group of patients undergoing PCI, measure (5) indicates a substantially higher mortality level compared to measure (1).

Comparing the PCI groups, and CABG and valve surgery groups, a patient death is more likely to be picked up through an inpatient re-admission within 30 days for the latter. For a more complete picture of mortality after PCI, including information after home discharge is more important.

¹ Patients with last known disposition "acute transfer" are excluded.
² Patients without record linkage number (RLN = ID that identifies patients across encounters) are excluded.
³ Out-Of-State Patients and Patients Who Left Against Medical Advice are excluded.

After focus group discussions and feedback from cardiologists, cardiac surgeons, health economists and consumers regarding a number of different possible outcomes, we chose 4 outcomes for the CCSIP reports:

Mortality

The mortality outcome for the CCSIP reports is based on death after the surgery admission or during connected subsequent acute care admissions or during any re-encounters within 30 days of surgery (see measure (3) in section Which mortality outcomes were considered for the CCSIP reports? above). This measure comes close to the STS measure "operative mortality":

" Operative Mortality includes: (1) all deaths, regardless of cause, occurring during the hospitalization in which the operation was performed, even if after 30 days (including patients transferred to other acute care facilities); and (2) all deaths, regardless of cause, occurring after discharge from the hospital, but before the end of the thirtieth postoperative day."
https://www.sts.org/sites/default/files/ACSD_TrainingManualV2-9_October2019(2).pdf, page 479

The mortality measure presented here does not include all deaths post-discharge outside an acute care setting: only deaths that occured in an acute care hospital or in an emergency room are included. Furthermore, procedures on California non-resident patients and patients who left against medical advice are excluded.

Patients undergoing surgery after Dec 1 of the last available year are excluded as it is not possible to follow these patients for 30 days post surgery.

The decision of using this outcome was driven by a) feedback we received after the release of the 2012 CCSIP results; b) its improvement in mortality reporting over measures (1) and (2), c) the ability to include all patients using outcome imputation for a small number of patients, and d) the characteristics of the model that was used to risk-adjust patient data (see section How were hospital-level outcomes compared?).

Multiple Adverse Cardiovascular and Cerebral Events (MACCE)

The MACCE outcome includes additional adverse events that might occur in the index admission or subsequent connected acute admissions. The MACCE outcome considers the following 4 adverse events:

1. death during the index admission or a subsequent connected admission;
2. post-op stroke that newly occurred during the index admission or stroke during a connected admission.
3. post-op acute MI that newly occurred during the index admission (POA=N) or acute MI during a connected re-admission;
4. Unplanned re-intervention:
   • for PCIs: Unplanned re-intervention during the index admission that occurred greater than 30 days or later after the date of the index PCI counting the date of the index PCI as day 1, or during a connected non-elective acute admission.
   • For isolated CABG, TAVR, isolated Surgical AV-Replacement, MV-Replacement, MV-Repair with or without CABG surgery: Re-intervention during the index admission after the date of the surgery or during a connected acute admission.
   Re-intervention events considered are cardiac procedures (prior October 1, 2015: 35.xx, 36.xx, 37.xx, 38.xx; October 1, 2015 or later: 02xxxxx) that are performed using extracorporeal circulation auxiliary to open heart surgery (ECC), or a CABG without use of ECC, or a TAVR, or a PCI.

MACCE and Contributing Adverse Events, California, 2020-2021
All contributing adverse events are based on the surgery/intervention admission and connected acute inpatient admissions.
Surgery / Intervention	Volume	(1) MACCE		(2) Mortality		(3) Post-OP Stroke		(4) Post-OP Myocardial Infarction		(5) Re-Intervention
		N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events
Isolated CABG	22,632	1,062	4.7	493	2.2	311	1.4	150	0.7	209	0.9
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair	7,386	351	4.8	185	2.5	153	2.1	12	0.2	28	0.4
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair with CABG	3,271	282	8.6	171	5.2	100	3.1	16	0.5	19	0.6
TAVR	15,380	396	2.6	139	0.9	222	1.4	23	0.2	47	0.3
PCI with ACS	55,587	3,391	6.1	2,570	4.6	498	0.9	103	0.2	253	0.5
PCI without ACS	40,467	1,478	3.7	551	1.4	161	0.4	978	2.4	69	0.2

Multiple Adverse Cardiovascular and Cerebral Events within 90 Days (MACCE-90)

In addition to the admissions considered for the MACCE outcome, the MACCE-90 outcome also checks inpatient re-admissions within 90 days of the index intervention / surgery. The MACCE-90 outcome considers the following 4 adverse events:

1. death during the index admission, a subsequent connected admission or within 90 days captured from inpatient or emergency department encounters;
2. post-op stroke that newly occurred during the index admission, stroke during a connected admission, or stroke during a re-admission within 90 days of the index procedure.
3. post-op acute MI that newly occurred during the index admission, acute MI during a connected re-admission or acute MI during a re-admission within 90 days of the index procedure;
4. Unplanned re-intervention:
   • for PCIs: Unplanned re-intervention during the index admission that occurred greater than 30 days or later after the date of the index PCI counting the date of the index PCI as day 1, or during a connected non-elective acute admission, or during a non-elective inpatient re-admission or ED encounter within 90 days of the index PCI.
   • For isolated CABG, TAVR, isolated Surgical AV-Replacement, MV-Replacement, MV-Repair with or without CABG surgery: Re-intervention during the index admission after the date of the surgery or during a connected acute admission, or during an inpatient re-admission, ambulatory surgery or ED encounter within 90 days of the index surgery.
   Re-intervention events considered are cardiac procedures (prior October 1, 2015: 35.xx, 36.xx, 37.xx, 38.xx; October 1, 2015 or later: 02xxxxx) that are performed using extracorporeal circulation auxiliary to open heart surgery (ECC), or a CABG without use of ECC, or a TAVR, or a PCI.

As the MACCE-90 outcome requires a 90-day follow-up period, surgeries that occurred after October 2 of the last available year are excluded. Also excluded are surgeries of out-of state residents and patients who left against medical advice.

The tables below show MACCE events occuring within 30 and 90 days of the index procedure.

MACCE-30 and Contributing Adverse Events, California, 2020-2021
All contributing adverse events are based on the surgery/intervention admission, connected acute inpatient admissions and re-admissions within 30 days of the index procedure.
Surgery / Intervention	Volume	(1) MACCE-30		(2) Mortality		(3) Post-OP Stroke		(4) Post-OP Myocardial Infarction		(5) Re-Intervention
		N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events
Isolated CABG	21,455	1,280	6.0	520	2.4	382	1.8	209	1.0	242	1.1
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair	6,922	422	6.1	189	2.7	177	2.6	28	0.4	32	0.5
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair with CABG	3,108	314	10.1	168	5.4	112	3.6	24	0.8	22	0.7
TAVR	14,534	602	4.1	199	1.4	294	2.0	67	0.5	97	0.7
PCI with ACS	52,157	4,393	8.4	2,754	5.3	666	1.3	93	0.2	937	1.8
PCI without ACS	38,352	2,336	6.1	657	1.7	235	0.6	1,387	3.6	373	1.0

MACCE-90 and Contributing Adverse Events, California, 2020-2021
All contributing adverse events are based on the surgery/intervention admission, connected acute inpatient admissions and re-admissions within 90 days of the index procedure.
Surgery / Intervention	Volume	(1) MACCE-90		(2) Mortality		(3) Post-OP Stroke		(4) Post-OP Myocardial Infarction		(5) Re-Intervention
		N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events	N Events	% Events
Isolated CABG	19,451	1,444	7.4	587	3.0	401	2.1	250	1.3	342	1.8
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair	6,304	439	7.0	205	3.3	169	2.7	35	0.6	46	0.7
Isolated SAV-Replacement, SMV-Replacement, SMV-Repair with CABG	2,855	343	12.0	196	6.9	110	3.9	31	1.1	30	1.1
TAVR	13,170	783	5.9	295	2.2	314	2.4	95	0.7	164	1.2
PCI with ACS	47,504	4,929	10.4	2,897	6.1	737	1.6	87	0.2	1,353	2.8
PCI without ACS	35,108	2,779	7.9	805	2.3	316	0.9	1,526	4.3	586	1.7

For the purpose of comparing hospital-leve outcome rates, due to differences in the mix of patients hospitals treat, it is not appropriate to calculate adverse outcome rates based on the observed number of adverse events and the number of patients undergoing a surgery or intervention. Hospitals that treat a predominantly relatively "healthy" mix of patients would look better compared to those hospitals whose patients are sicker.

This issue is addressed by fitting a statistical model for the adverse outcome and then using this statistical model to assess each hospital's patient mix. A risk-adjusted rate is then generated by multiplying the observed California statewide rate with the ratio of Observed to Expected adverse events. The models used for the CCSIP reports are included as links in the Methods section.

The table below lists the risk factors included in the models and their definition. Note that not all models include all the risk factors shown.

As ambulatory surgery and emergency department records do not include a POA indicator, unless noted otherwise, any diagnoses of comorbid conditions in an ambulatory surgery or emergency department record were considered present at admission.

Each table shows all hospitals with a procedure volume of at least 25 procedures for which the outcome can be determined. Children's hospitals are excluded. Note that this implies that some hospitals might only be listed for some of the surgeries and interventions considered for the CCSIP reports.

The current hospital name and the county where the hospital is located as reported by a hospital to HCAI (https://data.chhs.ca.gov/dataset/licensed-healthcare-facility-listing) is shown in the first two columns.

The third column Volume shows the volume of procedures performed at the hospital. By clicking on the column header, it is possible to sort the table in ascending order by volume; by clicking on the column header twice, it is possible to sort the table in descending order by volume. In addition, when hovering over the "Volume" column, the volume percentile is shown.

The fourth column Observed Events shows the number of adverse events. Clicking on the column header once sorts the table in ascending order by the number of adverse events; clicking on the column header twice sorts the table in descending order by the number of adverse events. Hovering over the "Observed Events" column displays a text box with the column header and percentile of adverse events.

The column Event % shows the un-adjusted adverse outcome rate. Clicking on the column header once sorts the table in ascending order by the adverse outcome rate; clicking on the column header twice sorts the table in descending order by the adverse outcome rate. Hovering over the "Events %" column displays a text box with the column header and percentile of the adverse outcome rate. Note that the percentile for the adverse outcome rate can be quite different from the risk-adjusted even rate if a hospital's patient mix differs considerably from the averageCalifornia patient.
For the MACCE and MACCE-90 outcomes, clicking on the number of observed events opens a mod with details on the composition of the adverse events for each year available.

The column Expected Event % shows the expected adverse outcome rate. This column allows us to assess how much a hospital's patient mix differs from the average California patient. A higher expected adverse outcome rate indicates that the average patient seen at this hospital was sicker than the average California patient. A lower expected adverse outcome rate indicates that the average patient seen at this hospital was healthier than the average California patient. Clicking on the column header once sorts the table in ascending order by the expected adverse outcome rate; clicking on the column header twice sorts the table in descending order by the expected adverse outcome rate. Hovering over the "Expected Event %" column displays a text box with the column header and percentile of the expected adverse outcome rate.

The column Risk-Adjusted Event % shows the risk-adjusted adverse outcome rate that is appropriate to compare adverse outcomes across hospitals. Clicking on the column header once sorts the table in ascending order by the risk-adjusted adverse outcome rate; clicking on the column header twice sorts the table in descending order by the risk-adjusted adverse outcome rate. Hovering over the "Risk-Adjusted Event %" column displays a text box with the column header and percentile of the risk-adjusted adverse outcome rate.

The last two table columns Lower 95% CL or RA % and Upper 95% CL of RA % provide confidence limits for for the risk-adjusted adverse outcome rate. These confidence limits might indicate that a hospital had significantly fewer adverse events than the California average hospital if the upper limit is lower than the observed California adverse outcome rate; they might indicate that a hospital had significantly more adverse events than the California average hospital if the lower limit is higher than the observed California adverse outcome rate. As all other columns in this table sorting by the columns' contents is enabled through clicking on the column headers.

If the lower 95% confidence limit for the risk-adjusted adverse outcome rate exceeds the California reference rate, the risk-adjusted adverse outcome rate is considered statistically significantly higher than the California average rate at the 5% confidence level. If the upper 95% confidence limit for the risk-adjusted adverse outcome rate is lower the California reference rate, the risk-adjusted adverse outcome rate is considered statistically significantly lower than the California average rate at the 5% confidence level. Statistically significant risk-adjusted adverse outcome rates are displayed bolded.

It is important to note that by chance alone we can expect 5% of hospitals to have a statistically significant adverse outcome rate. Putting current performance in the context of past performance, provides a more complete picture. For this reason, clicking on a hospital's name in the table gives access to trend charts that show risk-adjusted adverse outcome rates over time. The overlay that opens upon clicking on a hospital name shows the trend in the risk-adjusted rate from 1999 to 2021. A second chart shows the risk-adjusted rate with its confidence limits in a comparison chart for the most recent two years. for more details on these charts, refer to the next section in this document.

For the MACCE and MACCE-90 outcomes, clicking on the number of observed events opens a mod with details on the composition of the adverse events for each year available.

When clicking on the name of a hospital in a the Volume of all Cardiac Procedures table, two charts are displayed. The top chart shows the volume trend for the selected hospital from 1999 to 2021 for the following cardiac procedures:

• Isolated CABG,
• Isolated Surgical or Transcatheter AV-Replacement, MV-Replacement, MV-Repair with or without CABG
• Other Cardiac Surgery
• PCI with ACS, and
• PCI without ACS

by discharge year. The bottom chart shows the volume trend in a bar format that allows tracking the overall cardiac procedure volume as well and shows the actual number of procedures of each type on each bar. Note that the volume charts are only available for hospitals that had a minimum of 25 surgeries/interventions across the included years for at least one of the surgery / intervention groups.

When clicking on the name of a hospital in a the Volume of all Valve Procedures table, two charts are displayed.

The top chart shows a stacked bar chart with the volume trend for the selected hospital from 1999 to 2021 for the following valve procedures:

• Isolated Surgical AV-Replacement
• Isolated Surgical AV-Replacement with CABG,
• Isolated Transcatheter AV-Replacement
• Isolated Surgical MV-Replacement,
• Isolated Surgical MV-Replacement with CABG,
• Isolated Transcatheter MV-Replacement,
• Isolated Surgical MV-Repair,
• Isolated Surgical MV-Repair with CABG,
• Isolated Transcatheter MV-Repair,
• Multiple Surgical Valves including AV-Replacement, MV-Replacement and/or MV-Repair without CABG,
• Multiple Surgical Valves including AV-Replacement, MV-Replacement and/or MV-Repair with CABG,
• Other Surgical Valve Procedures,
• Other Transcatheter Valve Procedures.

Note that the group of other valve procedures also includes Surgical AV-Replacement, Surgical MV-Replacement, and Surgical MV-Repair procedures that were excluded from the isolated valves groups or isolated valves + CABG groups due to the presence of an exclusionary condition.

The bottom chart displays the volume trend as line chart with one line corresponding to each of the valve procedures outlined above. The volume charts are only available for hospitals that had a minimum of 25 procedures in the included years for at least one of the valve surgery groups.

In each chart the legend at the bottom can be used to toggle the display of the corresponding procedure in the chart.

When clicking on the name of a hospital in a the Mortality, MACCE and MACCE-90 tables, two charts are displayed.

The top chart displays the trend in the risk-adjusted adverse outcome rate from 1999 to 2021 for the hospital selected. Yearly rates are all standardized to the 2020-2021 California observed adverse outcome percentage, in other words, all rates can be compared over time as they are adjusted for case-mix. The band marked in light blue corresponds to a 95% confidence band around the hospital's risk-adjusted adverse outcome rate. If the bottom edge of this band exceeds the dashed reference line representing the California 2020-2021 levels of adverse outcomes, adverse outcome rates were higher during that year compared to the adverse outcomes experienced by the average California patient in 2020-2021. If the top edge of this band is below the dashed reference line representing the California 2020-2021 levels of adverse outcomes, adverse outcome rates were lower during that year compared to the adverse outcomes experienced by the average California patient in 2020-2021

Also shown in the top chart is a green dotted line. This line corresponds to the level of adverse outcomes in California for each year adjusted such that differences in the all California case-mix of patients are controlled for. If the dotted line is within the boundaries of the confidence band for a year, the hospital's risk-adjusted outcome rate was no different from the adverse outcome rate observed for California for that year.

The bottom chart shows all California hospitals with a minimum of 25 procedures of the type indicated and their risk-adjusted adverse outcome rates for 2020-2021 Hospitals are sorted by their risk-adjusted rates. Each hospital is represented by a bar that is determined based on 95% confidence limits of the risk-adjusted adverse outcome rate. The bar pertaining to the hospital selected is highlighted in orange. Note that the width (confidence limits) of each bar is an indicator of a hospital's volume in that narrower bar widths pertain to hospitals with a larger volume of cases. Bars that appear grayed or light orange are truncated at the maximum allowed for the chart. In other words, for these hospitals, the upper confidence limit exceeds the maximum value that can be displayed in the chart. The dashed reference line in the chart corresponds to the statewide observed adverse outcome rate.

The volume tables for all outcomes might show a higher volumes due to any of the following reaons:

• Patients who left against medical advice are excluded from outcomes reporting.
• Non-California residents are excluded from outcomes reporting.
• If a patient undergoes the same surgery / intervention more than once during the same admission, only the first of surgery / intervention is included in the reporting of outcomes. For instance, if a patient has a PCI on Day 1, and then a second PCI on Day 5 of the same admission, only the first PCI for this patient is included in the outcomes tables, but all PCIs are reported in the volume table.

Furthermore, for the mortality outcome, volumes in the outome section are lower as we are unable to follow patients, who had their surgery after December 1, 2021 for 30 days. Surgeries / interventions that occurred after December 1, 2021 are excluded from the Mortality outcome tables.

For the MACCE-90 outcome, volumes in the outcome section are lower as we are unable to follow patients, who had their surgery after October 2, 2021 for 90 days. Surgeries / interventions that occurred after October 2, 2021 are excluded from the MACCE-90 outcome tables.

An alternative data source for the use of PCIs in California is the CathPCI Registry. We have received reports from several facilities included in the CCSIP reports noting under-reporting of their PCI volume on the CCSIP website by 2% to 40%. Unfortunately, it is impossible to obtain access to the CathPCI Registry, therefore we are unable to provide a more thorough analysis on these differentials.

Another source of PCI volume data are utilization data published by HCAI. The utilization databases can either by downloaded as Microsoft Excel spreadsheets (https://hcai.ca.gov/data-and-reports/healthcare-utilization/hospital-utilization/) or queried on-line through the HCAI Financial & Utilization Reports.

When comparing the number of PCIs reported by California hospitals in the utilization data, we found that the definition of a PCI appears to vary by hospital and by year. To better understand what is reported in the utilization data, we generated the following metrics:

1. count of the number of PCI procedure codes found in an inpatient record / ambulatory surgery record;
2. count of number of stents placed;
3. count of patients with at least one PCI procedure code for each day hospitalized / encounter day (CCSIP approach);

We were unable to observe a consistent trend for any of the metrics. For some hospitals, the reported utilization from year to year varied substantially, or all years showed more or fewer PCIs for all comparison metrics in the utilization data compared to the inpatient, ambulatory surgery and emergency department data, or we found relatively good agreement for several years on one metric. The table below shows some examples. Note that ICD-10 coding started on October 1, 2015 which might impact the reported data as well. The arrow in the last column indicates whether the utilization imply a lower or higher number of PCIs compared to all comparison metrics.

Unfortunately, the utilization data is not a mandated program and facility data can be submitted / updated up to three years following the reporting period which limits our ability to perform a rigorous comparison. Also, the above table shows that for PCIs, there seem to be some definitional issues affecting the reliability of reported volumes.

In order to allow inclusion of all eligible patients into the Mortality, MACCE and MACCE-90 tables, we used a process called outcome imputation. In our focus groups and discussions with hospitals a frequent problem was the difference in the patient volume reported by the hospital and the volume included in a summary table. For instance, for the mortality outcome adopted for the CCSIP reports, it is not possible to determine whether or not a patient experienced an adverse event if the last known disposition for this patient was transfer. We adopted the following approach to allow the full patient volume to be used in the summary tables:

1. Only patients with a known outcome were included for the purpose of generating the model that was used to obtain the number of expected adverse events.
2. All patients were included when counting the number of expected events at each hospital, even those patients with unknown outcome status.
3. All patients were included when counting the number of observed events at each hospital. A patient who experienced the adverse event contributed a count of 1 to the adverse event count; a patient who did not experience the adverse event contributed a count of 0 to the adverse event count; a patient with unknown outcome status contributed the average observed adverse event probability for the hospital and for the discharge year. For instance, if the average observed mortality rate for a hospital was 2.0%, each patient with unknown outcome status, contributed 0.02 to the adverse event count.

This process of outcome imputation does not affect the observed adverse outcome rate for a hospital. It does make the assumption that for those patients for whom adverse event status cannot be determined, average conditions with respect to the outcome at the hospital during the year applied.

The tables below show the extent to which hospitals are affected by this process. Overall outcome imputation had to be used for a small number of cases allowing us to include all eligible patients in the risk-adjusted tables.

* Only hospitals listed in outcome tables included.

* Only hospitals listed in outcome tables included.

* Only hospitals listed in outcome tables included.

In the HCAI databases each licensed inpatient, ambulatory surgery or emergency department facility is referenced by a 6-digit hospital ID. These hospital IDs are associated with a hospital's physical location. Therefore, if a hospital moves to a new location, its HCAI ID changes. Or, alternatively, if a hospital changes ownership at the same physical location, the HCAI hospital code remains the same, but the hospital name might change. The hospital names displayed on this website are based on the most recent name provided by HCAI for the 6-digit hospital ID in the HCAI DBs.

A list of current and historical licenses California hospitals and their HCAI hospital IDs, name and address are available on the HCAI website.

HCAI assigns facility IDs based on a hospital's physical location. Therefore, if a hospital moves to a different location, its HCAI ID changes. Unless we are aware of the change in location , the hospital might then appear twice in the CCSIP report. Also, the longitudinal view of volumes and outcomes over time is lost.

For the following facilities, we have ensured that reporting is uninterrupted:

• Valleycare Medical Center (HCAI ID 014050) and Valley Memorial Hospital (HCAI ID 010983) consolidated in 2007. CCSIP reports all data for these facilities under Valleycare Medical Center (014050).
• Scripps Mercy Hospital - Chula Vista (HCAI ID: 370658) performed cardiac surgeries and interventions through 2008. For years prior to 2009, data for Scripps Mercy Hospital - Chula Vista are included in data for Scripps Mercy Hospital (HCAI ID: 370744) are reported under Scripps Mercy Hospital.
• Kaiser Foundation Hospital - Orange County - Lakeview (HCAI ID: 301132) moved in 2012. The new facility is Kaiser Foundation Hospital - Orange County - Anaheim (HCAI ID: 304409). CCSIP reports all data for these facilities under Kaiser Fnd Hosp - Orange County - Anaheim.
• Cardiac surgeries and interventions for Kaiser Foundation Hospital - Orange County Anaheim (HCAI ID: 304409) and Kaiser Foundation Hospital - Orange County Irvine (HCAI ID: 304306) are reported under Kaiser Foundation Hospital - Orange County.
• Kaiser Foundation Hospital - Oakland (HCAI ID: 010856) moved in 2014. The new facility is Kaiser Foundation Hospital - Oakland/Richmond (HCAI ID: 014326). CCSIP reports all data for these facilities under Kaiser Fnd Hsp - Oakland/Richmond.
• Sutter Medical Center of Santa Rosa (HCAI ID: 490919) moved in 2014. The new facility is Sutter Santa Rosa Regional Hospital (HCAI ID: 494106). CCSIP reports all data for these facilities under Sutter Santa Rosa Reg Hsp.
• Kaiser Foundation Hospital - Redwood City (HCAI ID: 410804) moved to a new location in 2014 with necessitated a new HCAI ID (414139). CCSIP reports all data for these facilities under HCAI ID 414139.
• Kaiser Foundation Hospital - Hayward (HCAI ID: 010858) closed in 2015. Its data are reported under Kaiser Foundation Hospital - Hayward/Fremont (HCAI ID: 014132).
• Sutter Memorial Hospital (HCAI ID: 341052) closed in 2015 and was replaced by Sutter General Hospital (HCAI ID: 341051). CCSIP reports all data for these facilities under HCAI ID 341051.
• Cardiac surgeries performed at UC San Diego Health Hillcrest (HCAI ID: 370782) are all reported under UC San Diego Health La Jolla - Jacobs Medical Center & Sulpizio Cardiovascular Center (HCAI ID: 374141).
• Cardiac surgeries and interventions at all UCSF Medical Center locations are reported under UCSF Medical Center (HCAI ID: 381154).
• Alta Bates Summit - Alta Bates Campus (HCAI ID: 010739) reported surgeries through 2002 and PCIs through 2011. This data is included under Alta Bates Medical Center (HCAI ID: 010937).
• Scripps Green (HCAI ID: 371256) reported surgeries and interventions through 2015. In 2015, Scripps Health opened the Prebys Cardiovascular Institute in La Jolla. Data for Scripps Green are combined with data for Scripps Memorial Hospital - La Jolla (HCAI ID: 370771).
• California Pacific Medical Center - Pacific Campus (HCAI ID: 380929) reported cardiac surgeries and interventions through March 2019. California Pacific Medical Center - Van Ness Campus (HCAI ID: 384176) opened in March 2019. All cardiac surgeries and interventions are reported for these two facilities are reported under California Pacific Medical Center - Van Ness Campus.
• Palomar Health Downtown Campus (HCAI ID: 370755) closed on 06/30/3031. Reporting is consolidated under Palomar Medical Center (HCAI ID: 374382).

For the following facilities, a substantial number of surgeries and/or interventions was reported during the study period. However, none were reported for the most recent two years.

• Doctors Medical Center - San Pablo (HCAI ID: 070904) reported surgeries through 2006 and PCIs through 2015.
• Lancaster Hospital (HCAI ID: 190455) reported surgeries and interventions through 2010.

Starting from July 2017, CCSIP reports are available to heart team members, health care providers and other stakeholders, including consumers, based on registration and agreement to terms of use. Use the Logon link to register or logon if you have previously registered.

While it is not possible to determine without a more extensive analysis that outcomes after cardiac procedures might have been lower had COVID-19 not occurred, it is interesting to look at the number of patients affected by COVID-19 during the outcome observation period and outcome rates by COVID-19 status.

Following the recommended COVID-19 coding guidelines, records pertaining to patients with a positive COVID-19 diagnosis were determined based on ICD-10-CM code U07.1 which should be assigned whenever a patient tested positive for COVID-19 irrespective of whether the patient was symptomatic. As this code was first introduced on April 1, 2020, the table below is based on patients discharged on or after that date.

COVID-19 status is determined based on the same follow-up period that is used for the outcome.

Among patients undergoing isolated CABG, the percent of patients with a positive COVID-19 diagnosis was 0.7% diagnosed during the index admission, a connected admission or within 30 days of the index procedure. Mortality in patients without / with a COVID-19 diagnosis was 2.3% and 14.6% respectively. A similar increase in the outcome percentage occurred for the other cardiac procedures and outcomes considered.

The table shows the Cochran-Mantel-Haenszel that formally tests for a non-zero correlation between COVID-19 status and outcome. A statistically significant correlation is observed except for the Isolated SAV-Replacement, SMV-Replacement, SMV-Repair groups and the MACCE or MACCE-90 outcomes.

The Relative Risk (RR) estimates the possible impact of COVID-19 on the outcome. For instance, the mortality after Isolated CABG surgery increased by 4.2% based on the prevalence of COVID-19 in the study population.

Of course, these considerations a crude. This analysis does not show that COVID-19 was the cause for an increase in adverse outcomes. A more sophisticated analysis would include patient risk factors and competing causes of death. Further, the effect of COVID-19 might be underestimated as the code to capture COVID-19 status might not have been widely adopted by hospitals. The last 2 columns of the table below shows the number and percent of hospitals for which we found at least 1 COVID-19 diagnosis code in the source data.

We are encouraging you to share your opinions about the CCSIP website. Please direct any comments you might have to the project staff listed on the Contacts link in the navigation bar. Many of the updates that were put in place in October 2017, October 2019, April 2021 and January 2023 were a result of feedback of website users.

We are also asking for your financial support. Please go to www.casts.org to support CA-STS/CCSIP.

All analytics was performed in SAS V9.4. SAS is a registered trademark of SAS Institute Inc.

The setup of all searchable tables was made possible through the use of the datatables.net JavaScript library. The use of this JavaScript library is supported under the MIT license.

All charts for the CCSIP reports were generated with Highcharts, a JavaScript-based charting library that is made available free of charge for non-commercial users.