Monitoring quality of care in acute myocardial infarction patients using retrospective registry data

Monitoring quality of care in acute myocardial infarction patients using retrospective registry data Abstract Background Hospital-based registries provide a key contribution in assessing the quality of care in acute myocardial infarction (MI) patients, although some concern on selection bias of included cases has recently arisen. We investigated the feasibility of a retrospective, population-based registry of MIs in monitoring the quality of care. Methods We identified all the hospitalizations with a diagnosis of acute MI among 35–79 years old residents in the Varese province, Northern Italy, in 2007–2008. Information needed to define performance according to the American Heart Association set was extracted from hospital case histories. To characterize our approach, we focus on data completeness for critical event times and eligibility criteria, and on the analysis of ST-elevated MI (STEMI) patients according to received reperfusion treatment. Results Exact time of hospital admission and of percutaneous coronary angioplasty (PCI) procedure was available in 96% and 77% of MIs, with no difference between non-transferred (n = 1399) and inter-hospital transferred (n = 300) patients. Data completeness for eligibility to action/treatment criteria was >90% for each performance measure except statin prescription at discharge (76%). About 45% of STEMI experienced a delay in PCI-capable hospital arrival, and only one every three ST-elevated MI patients received primary PCI; these were more likely to be younger male cases with less comorbidities than un-treated patients. Conclusions Complementary to clinical registries, the retrospective population-based is a feasible approach which allows monitoring the entire pattern of care of all hospitalized MI patients independent of their clinical characteristics. quality indicators, hospital care, registries, acute myocardial infarction, integrated care pathway, selection bias, disease management Introduction Cardiovascular disease accounts for about one-third of deaths worldwide [1], with a total estimated cost in the USA of $312 billion, exceeding by about 40% the total cost for cancer diagnosis and treatment [2]. Not surprisingly, health administrators and insurances are interested in monitoring healthcare performance and outcomes through standardized indicators obtained from reliable data. Hospital-based registries of acute myocardial infarction (MI) patients are currently considered the reference standard in documenting how patients are treated relative to best practice recommendations outside the ideal world of clinical trials [3]. These registries have been set-up in the USA [4–6], in Europe and UK [7, 8] and the GRACE registry has extended the same approach in a multinational project enrolling MI patients in three different continents [9]. Some concern has arisen lately on the validity of these registries since selection bias; it has been reported that two registries included MI cases at lower risk than MI cases in the reference population [10–13]. Moreover, these registries often do not include in the performance reports the patients experiencing either a delay in hospital arrival or an inter-hospital transfer during the acute phase, making it difficult to monitor the entire process of acute care [8, 14]. These restrictions may constitute key limitations when these registries are used for public health planning [15] or for outcome research [16], calling for alternative approaches. Population-based registries provided an enormous contribution to cardiovascular diseases epidemiology. The methodological approach in these registries is to utilize consolidated administrative databases as sources of event identification and obtain clinical details via case note abstraction (‘cold-case pursuit [17]’). In an acute condition such as the myocardial infarction, the retrospective study design mitigates against selection bias since all hospitalized cases with a discharge diagnosis of MI are included [18] and it may provide longitudinal data on the patient. Challenging aspects to this approach are the completeness of data collection, since information on quality of care is retrospectively extracted from case records, as well as timing in final results availability. Recently, a population-based approach has been shown to be feasible to describe quality of cancer care [19]. In this methodological paper, we aim to evaluate the feasibility of monitoring the quality of care in hospitalized acute MI events using a population-based registry established in the Province of Varese, Northern Italy. We focus on the data completeness for critical event times and eligibility to action or treatment criteria. In addition, we characterize all patients being discharged with a diagnosis of ST-elevated MI (STEMI), including those experiencing a delay in hospital arrival or an intra-hospital transfer, to document the importance of monitoring the entire process of care. Methods Study setting and event identification The MI CAMUNI (CArdiovascular Monitoring Unit in Northern Italy) Register is a population-based registry of MI events occurred to 35–79-years-old residents in the Varese Province in 2007–2008, extending the WHO MONICA MI Register established in the nearby Brianza since 1985 [20]. The Varese Province is located in the Lombardia Region between Milan and the Swiss border. Since the clinical performance measures are referring to the in-hospital setting, for the purpose of the present report we focus on hospitalized cases. The data source for the in-hospital events was the Regional Hospital Discharge Diagnoses (HDD) database. Selection criterion was the presence of an ICD-9 410.xx (except 410.x2) as main discharge diagnosis. The HDD database is an administrative healthcare archive held mainly for hospital reimbursement, with no detailed clinical information on the patients. Being based upon a retrospective data collection, the register did not need a formal approval by an Institutional Review Board. Data collection Case reports of selected hospitalizations were reviewed by trained medical personnel to extract relevant information to assess performance: patient demographics, clinical features (symptom-onset time, first medical contact details, initial ECG, hospital arrival, out-of-hospital cardiac arrest, symptoms and cardiac markers), in-hospital management (aspirin at arrival, fibrinolysis, percutaneous coronary angioplasty (PCI), left ventricular systolic function evaluation), clinical reasons for not performing or delaying a treatment, previous history of MI and cardiovascular disease risk factors (smoking status, diabetes, hypertension and family history of MI), drug treatment at admission and at discharge. Approximate event time was collected when it was the only available information (i.e. time between symptoms onset and hospital arrival less than 12 h). The study personnel received a study manual and periodical training during the 12-month data collection period; a few meetings were dedicated to discuss controversial cases. During data collection, data quality checks were carried out to detect missing values and data inconsistency. All data violations were solved and missing values confirmed whenever information was not present in the case history. A final data quality assessment is available online [21]. Definition of acute MI events Hospitalizations due to non-acute (elective) interventions such as coronary angiography or coronary artery bypass grafting were excluded (n = 83). Consecutive hospitalizations on the same patient due to hospital transfer were considered as one acute MI event (n = 39, of which 17 within the first 24 h). Access to the Emergency Department of one hospital followed by transfer to another study hospital was considered as transferred event. Each event ends with either the patient’s death, transfer to another hospital for elective interventions or rehabilitation, or discharge to home. Definition of quality of care Quality of care was assessed according to the American Heart Association performance measures [22], which quantitatively detail the adherence to treatment guidelines during the acute phase of the event (aspirin administration at arrival, reperfusion treatment for STEMI patients), the hospital stay (LDL-cholesterol assessment, evaluation of left ventricular systolic function), and at discharge (secondary prevention: drug prescription, smoking cessation advice, patient’s referral to cardiac rehabilitation). Performance measures were expressed as either (i) proportion of patients treated according to guidelines relative to those eligible for action/treatment; or (ii) timeliness of reperfusion therapy (median time to fibrinolysis/PCI) among re-perfused STEMI patients. Each indicator requires to define the ‘eligible population’, i.e. those who should have been treated according to guidelines excluding subjects with specific contraindications to action/treatment (allergy for instance) or other non-medical reasons leading to ineligibility. The complete list of performance measures and eligibility criteria is provided as Supplementary Table 1. It is worth noting that since our approach allowed collecting data at all hospital facilities for transferred patients, there was no need to exclude these patients in calculating the quality of care measures, including aspirin treatment at arrival and door-to-needle and door-to-balloon times. Statistical analysis Data completeness was presented as the prevalence of missing data on (i) critical date/times, for transferred and not transferred events; and (ii) on eligibility criteria for each performance measure. In addition, we compared clinical and demographic characteristics of four categories of STEMI patients: late hospital arrival (>12 h of symptoms onset); arrival within 12 h but in a non-PCI capable hospital; arrival within 12 h to a PCI-capable hospital but not receiving primary PCI; arrival to a PCI-capable hospital within 12 h and receiving primary (i.e. within 24 h of hospital arrival) PCI. Categories have been defined following clinical eligibility to primary PCI treatment [22]. The order of presentation corresponds to an increasing likelihood of the patients to be represented by clinical registries, the first category being always excluded and the last satisfying inclusion to procedural registries of PCI patients. Major demographic and clinical characteristics were summarized as mean (standard deviation) or proportion. We tested the null hypothesis of no difference among the groups using an analysis of variance or a chi-square test for continuous and categorical variables, respectively. All the analyses were conducted using SAS software version 9.4. Results Characteristics of the study population From the Regional Hospital Discharge Diagnoses (HDD) database, we identified n = 1830 hospitalizations with a diagnosis of 410.xx at discharge in one of the 11 hospitals in the study area (one university hospital and four PCI referral facilities) occurred to residents in the Varese Province (Fig. 1). We were able to access 1821 of the 1830 hospital clinical records (99.5%), corresponding to 1699 acute MI events after the case history review (Fig. 1). The patients had an average age of 64.6 years (SD: 10.6 years); 72% of them were men; 17% (n = 300) were transferred from the Emergency Department of another hospital, 65% (n = 1098) had a discharge diagnosis of STEMI, and 3.7% (n = 63) died in hospital. Figure 1 View largeDownload slide Event identification from the Hospital Discharge Records database, data collection and definition of the number of acute MI events after considering in-hospital transfers at admission, including transfer from Emergency Department of another facility. The CAMUNI registry. Figure 1 View largeDownload slide Event identification from the Hospital Discharge Records database, data collection and definition of the number of acute MI events after considering in-hospital transfers at admission, including transfer from Emergency Department of another facility. The CAMUNI registry. Data completeness Table 1 reports the percentages of available data for the entire sample and for transferred and non-transferred patients. Exact time of first medical contact and of hospital admission was available in 82% and 96% of patients, respectively, with no difference according to transfer status (P-values >0.5). For transferred events it was more difficult to establish the exact timing of symptoms onset (67% vs. 74% in non-transferred patients; P = 0.02). However, information on presentation time within 12 h of symptoms onset was available for 98% of all events, regardless of transfer status. The exact time of balloon inflation was available in 77% of STEMI patients that underwent PCI, with no difference according to transfer status (P = 0.3). Nevertheless, it was possible to establish whether a PCI occurred in the first 24 h of hospital arrival in all the patients, regardless of inter-hospital transfer. Finally, exact time of fibrinolysis had the highest prevalence of missing data (66%), especially so for treatment administered in the Emergency Department before inter-hospital transfer (56%: chi-square test P-value = 0.07). Table 1 Proportion of data availability (%) for critical event times, for all acute MI events and by type of hospital admission. The CAMUNI Registry   Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0      Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0    aExact time is defined as the presence in the case history of the date/time of action/procedure; approximate time is referred to the presence in the case history of enough information to establish at least whether an action/treatment happens within a certain amount of time. bTransferred patients: arrival at the emergency department of the first hospital. cAmong those receiving PCI during the first 24 h of hospital arrival. dChi-square test P-value for testing a difference in data completeness according to transfer status. Table 1 Proportion of data availability (%) for critical event times, for all acute MI events and by type of hospital admission. The CAMUNI Registry   Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0      Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0    aExact time is defined as the presence in the case history of the date/time of action/procedure; approximate time is referred to the presence in the case history of enough information to establish at least whether an action/treatment happens within a certain amount of time. bTransferred patients: arrival at the emergency department of the first hospital. cAmong those receiving PCI during the first 24 h of hospital arrival. dChi-square test P-value for testing a difference in data completeness according to transfer status. Table 2 details the proportion of missing data on eligibility criteria and on action/treatment for each performance measure. Missing data on the eligible criteria exceeded 10% for the ‘statin prescription at discharge’ measure, due to 23% of events with no LDL-cholesterol assessment during the hospital stay. More than 5% of missing data were observed for left ventricular systolic dysfunction (9.4%), lipid-lowering therapy as pre-arrival medication (8.3%), smoking status (7.4%) and Warfarin/Coumadin as a pre-arrival medication (7.4%; performance measure ‘Aspirin at arrival’). The proportion of eligible MI events depends on the population of interest, ranging from 18% for ACE-inhibitor prescription for patients with left ventricular systolic dysfunction, to 100% for left ventricular function evaluation during the hospital stay. Finally, among eligible, the proportion of missing data on action/treatment was below 5% for all the indicators. Table 2 Data completeness of eligibility criteria and action/treatment for each performance measure. The CAMUNI Registry Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  NA: not applicable (no patient’s based eligibility criteria). LVSD, left ventricular systolic dysfunction (documented left ventricular ejection fraction < 40%). ACE-I, angiotensin-converting-enzyme inhibitor. aThis performance measure is not relevant for in-hospital deaths during the first 24 h (n = 32). bThis performance measure is relevant for STEMI events only (n = 1098). cThis performance measure is not relevant for in-hospital deaths (n = 63). See Table 1 in the supplementary material for a complete description of patient’s based exclusion criteria for each performance measure. dPercentage of MI events with missing data on eligibility criteria, on the total number of acute MI events (column 1). ePercentage of MI events with missing data on action/treatment, on the total number of eligible acute MI events (column 3). Table 2 Data completeness of eligibility criteria and action/treatment for each performance measure. The CAMUNI Registry Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  NA: not applicable (no patient’s based eligibility criteria). LVSD, left ventricular systolic dysfunction (documented left ventricular ejection fraction < 40%). ACE-I, angiotensin-converting-enzyme inhibitor. aThis performance measure is not relevant for in-hospital deaths during the first 24 h (n = 32). bThis performance measure is relevant for STEMI events only (n = 1098). cThis performance measure is not relevant for in-hospital deaths (n = 63). See Table 1 in the supplementary material for a complete description of patient’s based exclusion criteria for each performance measure. dPercentage of MI events with missing data on eligibility criteria, on the total number of acute MI events (column 1). ePercentage of MI events with missing data on action/treatment, on the total number of eligible acute MI events (column 3). Comparison of STEMI patients groups Patients’ flow-chart and likelihood of receiving on-site primary PCI are presented in Fig. 2. According to guidelines [23], STEMI patients with a time from symptom onset to hospital arrival of less than 12 h and who presented to a PCI-capable facility (n = 600, 55% of all STEMI) are eligible for on-site primary PCI. Of these, 310 patients (28% of initial STEMI population) actually received treatment; the median door-to-balloon time was 78 min (interquartile range: 50, 136 min), and 58% were treated within 90 min. Patients treated with on-site primary PCI were 4 years younger, more likely to be men and smokers, and less likely to have a history of MI, diabetes and hypertension than those who did not receive any primary reperfusion therapy (Table 3, columns (3) and (4); all P-values <0.05). In-hospital mortality was almost 4-fold larger among un-treated (8.5%) than among treated patients (2.3%; P = 0.001). Compared to patients experiencing either a delay in hospital arrival (column 1) or admitted to a non-PCI capable hospital (column 2), STEMI patients receiving primary PCI were younger (P = 0.01), less likely to have diabetes (P < 0.0001) and more likely to have used a pre-hospital medical service (P < 0.0001). Patients first admitted to a non-PCI capable hospital had twice the in-hospital mortality rate (5.1%) than treated patients. Table 3 Clinical characteristics at admission and in-hospital mortality of ST-elevation MI patients according to time between symptoms onset and hospital arrival, type of hospital arrival, and reperfusion treatment. ST-elevation MI patients with information on time of symptom onset (n = 1082). The CAMUNI Registry   Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09    Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09  aExcluding those who had fibrinolysis (n = 83) and those with treatment contraindications (n = 6) bP-value testing the difference between (4) and (3). T-test and chi-square test for continuous and categorical variables, respectively cP-value testing the difference between (4), (1) and (2) [2df test]. F-test and chi-square test for continuous and categorical variables, respectively Table 3 Clinical characteristics at admission and in-hospital mortality of ST-elevation MI patients according to time between symptoms onset and hospital arrival, type of hospital arrival, and reperfusion treatment. ST-elevation MI patients with information on time of symptom onset (n = 1082). The CAMUNI Registry   Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09    Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09  aExcluding those who had fibrinolysis (n = 83) and those with treatment contraindications (n = 6) bP-value testing the difference between (4) and (3). T-test and chi-square test for continuous and categorical variables, respectively cP-value testing the difference between (4), (1) and (2) [2df test]. F-test and chi-square test for continuous and categorical variables, respectively Figure 2 View largeDownload slide Patients’ flow-chart and likelihood of receiving reperfusion treatment, and median door-to-balloon time (interquartile range) for STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset. The CAMUNI registry. Patients’ selection was made according to reference [22] for door-to-balloon time in patients receiving on-site PCI within the first 24 h of hospital arrival (primary PCI). ‘Door’ time is the arrival at the emergency department. *n = 16 MI events with no information on time of symptoms onset. Figure 2 View largeDownload slide Patients’ flow-chart and likelihood of receiving reperfusion treatment, and median door-to-balloon time (interquartile range) for STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset. The CAMUNI registry. Patients’ selection was made according to reference [22] for door-to-balloon time in patients receiving on-site PCI within the first 24 h of hospital arrival (primary PCI). ‘Door’ time is the arrival at the emergency department. *n = 16 MI events with no information on time of symptoms onset. Discussion In this paper, we document the feasibility and the importance of a population-based registry for monitoring the quality of care in hospitalized acute myocardial infarction patients, as defined by standard performance indicators [22]. Event selection was based on the main discharge diagnosis code 410.xx, which was shown to provide reliable data on acute myocardial infarction in Northern Italy; [24] we were able to review the case histories of 99% of selected cases. Data completeness in hospital-based registries is an issue whenever missing data are related to event severity [11]. In our population-based registry, we were able to establish most of the critical date/time, including presentation time and primary PCI execution, in virtually all the events and regardless of initial transfer status. One PCI-capable hospital began to report exact PCI time only from 2008 onwards, contributing to the 23% of missing date/time of PCI intervention; however missing data was independent of the patient’s characteristics including age, gender, presentation delay from symptom onset, prevalence of typical symptoms and in-hospital mortality (data not shown). Eligibility criteria are essential to carefully define the subset of patients who should have been treated according to guidelines and thus ultimately provide unbiased estimates of performance [22]. We were able to collect information on most of the eligibility criteria, as documented in the Supplementary Table 1. Although it is difficult to compare the prevalence of eligible patients in different contexts, the prevalence of smokers, aspirin allergy and left ventricular systolic dysfunction in our study was similar to other reports in Italy [25] and the USA [14]. Missing data on eligible criteria were mostly due to either lack of information on pre-arrival medication (for ASA at arrival and LDL-cholesterol assessment performance measures) or no risk evaluation during the hospital stay (no LDL-cholesterol assessment, no left ventricular systolic function evaluation). Patients may not receive an LDL-assessment because they are considered at high risk [26] and thus fulfill the treatment indication (83% prevalence of statin prescription at discharge in our registry in this group, compared to 90% in the group with LDL-cholesterol >100 mg/dl and 77% in patients with LDL-cholesterol <100 mg/dl). Implications Quality of care assessment as defined by core hospital performance measures is undertaken mostly within a hospital-based perspective which often excludes the reporting of patients transferred during the acute event phase, as well as of patients with a late hospital presentation [8, 14, 27]. In our registry, the median door-to-balloon time for STEMI patients admitted to a PCI-capable facility within 12 h of symptoms onset was 78 min, similar to what reported by procedural registries in the same period in Italy [25]. These patients were comparable to the average patient profile in the PCI procedures registry active in the Lombardia region during the same time span: mean age 63 years, 77% men, 15% with diabetes and 40% admission through the Emergency Medical System [28]. However, as documented by Fig. 2 and Table 3, they are the expression of a care pathway which selects only 28% of initial STEMI patients: any further improvement in door-to-balloon time will affect less than 1 patient out of 3. About 44% of STEMI patients, in particular older women with comorbidities and patients less likely to access to the pre-hospital medical service, experienced either a delay in hospital presentation or a first arrival to a non-PCI capable hospital. A further 18% of STEMI patients arrived to a PCI-capable hospital within 12 h of symptoms onset, but did not receive primary PCI. These were on average 4 years older; were more likely to have comorbidities (history of MI, diabetes, hypertension); and experienced an in-hospital mortality rate almost four times higher than treated patients. By design, clinical and procedural registries are not feasible to quantify nor to appropriately monitor the patients’ selection which affects the care process: in a systematic review of 129 published papers from 77 clinical registries, only 3% were aware of the possibility of selection bias [12]. The advantage of the retrospective approach is to allow a comprehensive evaluation of the entire acute care pathway, providing additional information that stakeholders and decision makers can use in an integrated disease management plan [29, 30]. The proportion of STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset can be an exemplification. First, the retrospective approach can be used to identify population subgroups (the elderly, women) to be targeted by specific interventions to increase symptoms awareness and facilitate the use of pre-hospital medical service. Then, the process indicator can be monitored by extending the registry over time. Finally, through record linkage to external data such as the routine mortality database or the hospital discharge records, it is possible to assess the efficacy of the intervention in the reduction of re-hospitalizations and mortality. Current limitations and future improvements in study design The study design might be improved by considering a weighted sampling of cases to reduce duration and costs of data collection, by improving the timeliness of findings availability. Data completeness depends on case histories, and might vary in other settings. Based on our previous experience on epidemiological registries, we fixed an upper age limit at 79 years due to the fact that collecting reliable data on older patients is very challenging within our system. Moreover, we limited our registry to residents in the Varese Province only (i.e. non-residents cases were excluded). Residency is a standard inclusion criterion for many epidemiologic registries. Based on our data, non-residents cases were on average younger than residents (63.6 vs. 64.9 years old at hospital admission) and many performance measures are associated with the patient’s age [31, 32]. Therefore this choice is not recommended in future studies but it is not likely to have affected our study implications due to the study design. Conclusions In conclusion, this study documented the feasibility and importance of a population-based registry to assess the quality of hospital management in acute MI patients. Complementary to clinical registries, the population-based approach allows monitoring of the entire acute care process, providing incremental valuable information for public health planning and resources allocation strategies. Supplementary material Supplementary material is available at International Journal for Quality in Health Care online. Acknowledgments The authors are grateful to Dr. A. Borsani and Dr. M. Bonzini for developing and testing the case report form, as well as for coordinating the data collection activities. We thank Drs. S. Mombelli, M. Casà N. Facchinetti, S. Landone, M. Conti, R. Corrao and S. Colombo for review of case histories, data collection and data checking; and Dr. M. de Biasi for data quality assessment. We also thank Dr. M. Marzegalli, Cardiology Unit of the San Carlo Hospital, Milan, and Dr. O. Leoni, Agenzia di Tutela della Salute dell’Insubria, for valuable comments on earlier drafts of this manuscript. Funding This work was supported by the Italian Ministry of Health [Grant Ricerca Finalizzata Programma Strategico-2007-2-634753]. G.V. work was supported by a grant from the Health Administration of Lombardia Region [Grant 10800/2009] as part of the ‘Osservatorio Regionale Lombardo per le malattie cardiovascolari’. Conflict of interest statement None. References 1 Lozano R, Naghavi M, Foreman K et al.  . Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet  2012; 380: 2095– 128. Google Scholar CrossRef Search ADS PubMed  2 Writing Group Members. Mozaffarian D, Benjamin EJ, Go AS et al.  . American Heart Association Statistics Committee; Stroke Statistics Subcommittee. Heart Disease and Stroke Statistics-2016 update: a report from the American Heart Association. Circulation  2016; 133: e38– 360. Google Scholar CrossRef Search ADS PubMed  3 Gitt AK, Bueno H, Danchin N et al.  . The role of cardiac registries in evidence-based medicine. Eur Hearth J  2010; 31: 525– 29. Google Scholar CrossRef Search ADS   4 Gibson CM, Pride YB, Frederick PD et al.  . Trends in reperfusion strategies, door-to-needle and door-to-balloon times, and in-hospital mortality among patients with ST-segment elevation myocardial infarction enrolled in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J  2008; 156: 1035– 44. Google Scholar CrossRef Search ADS PubMed  5 Laskey W, Spence N, Zhao X et al.  . Regional differences in quality of care and outcomes for the treatment of acute coronary syndromes. an analysis from the get with the Guidelines Coronary Artery Disease Program. Crit Pathways Cardiol  2010; 9: 1– 7. Google Scholar CrossRef Search ADS   6 Vogeli C, Kang R, Landrum MB et al.  . Quality of care provided to individual patients in US hospitals. Results from an analysis of National Hospital Quality Alliance Data. Med Care  2009; 47: 591– 9. Google Scholar CrossRef Search ADS PubMed  7 Widimsky P, Wijns W, Fajadet J et al.  . European Association for Percutaneous Cardiovascular Interventions. Reperfusion therapy for ST elevation acute myocardial infarction in Europe: description of the current situation in 30 countries. Eur Heart J  2010; 31: 943– 57. Google Scholar CrossRef Search ADS PubMed  8 Balzi D, Di Bari M, Barchielli A et al.  . Should we improve the management of NSTEMI? Results from the population-based ‘acute myocardial infarction in Florence 2’ (AMI-Florence 2) registry. Intern Emerg Med  2013; 8: 725– 33. Google Scholar CrossRef Search ADS PubMed  9 GRACE investigators. Rationale and design of the GRACE (Global Registry of Acute Coronary Events) Project: a multinational registry of patients hospitalized with acute coronary syndromes. Am Heart J  2001; 141: 190– 9. CrossRef Search ADS PubMed  10 Krumholz HM. Registries and selection bias: the need for accountability. Circ Cardiovasc Qual Outcomes  2009; 2: 517– 8. Google Scholar CrossRef Search ADS PubMed  11 McNamara RL. Cardiovascular registry research comes of age. Heart  2010; 96: 908– 10. Google Scholar CrossRef Search ADS PubMed  12 Ferreira-Gonzalez I, Marsal JR, Mitjavila F et al.  . Patient registries of acute coronary syndrome. Assessing or biasing the clinical real world data? Circ Cardiovasc Qual Outcomes  2009; 2: 540– 7. Google Scholar CrossRef Search ADS PubMed  13 Rosvall M, Ohlsson H, Hansen O et al.  . Auditing patient registration in the Swedish quality register for acute coronary syndrome. Scand J Public Health  2010; 38: 533– 40. Google Scholar CrossRef Search ADS PubMed  14 Bernheim S, Wang Y, Bradley EH et al.  . Who is missing from the measure? Trends in proportion and treatment of patients potentially excluded from publicly-reported quality measures. Am Heart J  2012; 160: 943– 50. Google Scholar CrossRef Search ADS   15 Larsson S, Lawyer P for the Boston Consulting Group. Improving health care value. The case for disease registries. [https://www.bcgperspectives.com/content/articles/health_care_payers_providers_biopharma_improving_health_care_value_disease_registries/] (December 11, 2017, date last accessed). 16 Roger VL. Outcome research and epidemiology. The synergy between public health and clinical practice. Circ Cardiovasc Qual Outcomes  2011; 4: 257– 9. Google Scholar CrossRef Search ADS PubMed  17 Tunstall-Pedoe H. Problems with criteria and quality control in the registration of coronary events in the MONICA study. Acta Med Scand Suppl  1988; 728: 17– 25. Google Scholar PubMed  18 Brieger D, Aliprandi-Costa B. Developments in procedural and disease registries: a focus on coronary artery disease. Curr Opin Cardiol  2013; 28: 405– 10. Google Scholar CrossRef Search ADS PubMed  19 Caldarella A, Amunni G, Angiolini C et al.  . Feasibility of evaluating quality cancer care using registry data and electronic health records: a population-based study. Int J Qual Health Care  2012; 24: 411– 8. Google Scholar CrossRef Search ADS PubMed  20 Veronesi G, Ferrario MM, Chambless LE et al.  . The effect of revascularization procedures on myocardial infarction incidence rates and time trends: The MONICA-Brianza and CAMUNI MI registries in Northern Italy. Ann Epidemiol  2012; 22: 547– 53. Google Scholar CrossRef Search ADS PubMed  21 Veronesi G, De Biasi M, Ferrario MM Registry of hospitalized Acute Myocardial infarction in the Varese Province: Data Quality Assessment. Available at: http://www4.uninsubria.it/on-line/home/naviga-per-tema/ricerca-scientifica/centri-di-ricerca/centro-di-ricerca-in-epidemiologia-e-medicina-preventiva-epimed/documento307171.html (December 11, 2017, date last accessed). 22 Writing Committee to Develop Performance Measures for ST-Elevation and Cardiology/American Heart Association Task Force on Performance Measures. ACC/AHA 2008 performance measures for adults with ST-Elevation and Non ST-Elevation Myocardial Infarction. A report of the American College of Cardiology/American Heart Association Task Force on Performance Measures. Circulation  2008; 118: 2596– 648. CrossRef Search ADS PubMed  23 Piepoli MF, Hoes AW, Agewall S et al.  . 2016 European Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J  2016; 37: 2315– 81. Google Scholar CrossRef Search ADS PubMed  24 Brocco S, Fedeli U, Schievano E et al.  . Effect of the new diagnostic criteria for ST-elevation and non-ST elevation acute myocardial infarction on 4-year hospitalization: an analysis of hospital discharge records in the Veneto Region. J Cardiovasc Med  2006; 7: 45– 50. Google Scholar CrossRef Search ADS   25 Manari A, Ortolani P, Guastaroba P et al.  . Clincal impact of an inter-hospital transfer strategy in patients with ST-elevation myocardial infarction undergoing primary angioplasty: the Emilia-Romagna ST-segment elevation acute myocardial infarction network. Eur Hearth J  2008; 29: 1834– 42. Google Scholar CrossRef Search ADS   26 Javed U, Deedwania PC, Bhatt DL et al.  . Use of intensive lipid-lowering therapy in patients hospitalized with acute coronary syndrome: an analysis of 65 396 hospitalizations from 334 hospitals participating in Get With The Guidelines (GWTG). Am Heart J  2011; 161: 418– 24. Google Scholar CrossRef Search ADS PubMed  27 Fonarow GC, Gregory T, Driskill M et al.  . Hospital certification for optimizing cardiovascular disease and stroke quality of care and outcomes. Circulation  2010; 122: 2459– 69. Google Scholar CrossRef Search ADS PubMed  28 Martinoni A, De Servi S, Boschetti E et al.  . Lombardima Study Group. Importance and limits of pre-hospital electrocardiogram in patients with ST elevation myocardial infarction undergoing percutaneous coronary angioplasty. Eur J Cardiovasc Prev Rehabil  2011; 18: 526– 32. Google Scholar CrossRef Search ADS PubMed  29 Romeyke T, Stummer H. Clinical pathways as instruments for risk and cost management in hospitals - a discussion paper. Glob J Health Sci  2012; 4: 50– 9. Google Scholar CrossRef Search ADS PubMed  30 Bufalino VJ, Masoudi FA, Stranne SK et al.  . for the American Heart Association Advocacy Coordinating Committee. The American Heart Association’s recommendations for expanding the applications of existing and future clinical registries: a policy statement from the American Heart Association. Circulation  2011; 123: 2167– 79. Google Scholar CrossRef Search ADS PubMed  31 Peterson ED, Shah BR, Parsons L et al.  . Trends in quality of care for patients with acute myocardial infarction in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J  2008; 156: 1045– 55. Google Scholar CrossRef Search ADS PubMed  32 Schoenenberger AW, Radovanovic D, Stauffer JC et al.  . Acute Myocardial Infarction in Switzerland Plus Investigators. Age-related differences in the use of guideline-recommended medical and interventional therapies for acute coronary syndromes: a cohort study. J Am Geriatr Soc  2008; 56: 510– 6. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal for Quality in Health Care Oxford University Press

Monitoring quality of care in acute myocardial infarction patients using retrospective registry data

Loading next page...
 
/lp/ou_press/monitoring-quality-of-care-in-acute-myocardial-infarction-patients-RSGnohPea4
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1353-4505
eISSN
1464-3677
D.O.I.
10.1093/intqhc/mzy025
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Hospital-based registries provide a key contribution in assessing the quality of care in acute myocardial infarction (MI) patients, although some concern on selection bias of included cases has recently arisen. We investigated the feasibility of a retrospective, population-based registry of MIs in monitoring the quality of care. Methods We identified all the hospitalizations with a diagnosis of acute MI among 35–79 years old residents in the Varese province, Northern Italy, in 2007–2008. Information needed to define performance according to the American Heart Association set was extracted from hospital case histories. To characterize our approach, we focus on data completeness for critical event times and eligibility criteria, and on the analysis of ST-elevated MI (STEMI) patients according to received reperfusion treatment. Results Exact time of hospital admission and of percutaneous coronary angioplasty (PCI) procedure was available in 96% and 77% of MIs, with no difference between non-transferred (n = 1399) and inter-hospital transferred (n = 300) patients. Data completeness for eligibility to action/treatment criteria was >90% for each performance measure except statin prescription at discharge (76%). About 45% of STEMI experienced a delay in PCI-capable hospital arrival, and only one every three ST-elevated MI patients received primary PCI; these were more likely to be younger male cases with less comorbidities than un-treated patients. Conclusions Complementary to clinical registries, the retrospective population-based is a feasible approach which allows monitoring the entire pattern of care of all hospitalized MI patients independent of their clinical characteristics. quality indicators, hospital care, registries, acute myocardial infarction, integrated care pathway, selection bias, disease management Introduction Cardiovascular disease accounts for about one-third of deaths worldwide [1], with a total estimated cost in the USA of $312 billion, exceeding by about 40% the total cost for cancer diagnosis and treatment [2]. Not surprisingly, health administrators and insurances are interested in monitoring healthcare performance and outcomes through standardized indicators obtained from reliable data. Hospital-based registries of acute myocardial infarction (MI) patients are currently considered the reference standard in documenting how patients are treated relative to best practice recommendations outside the ideal world of clinical trials [3]. These registries have been set-up in the USA [4–6], in Europe and UK [7, 8] and the GRACE registry has extended the same approach in a multinational project enrolling MI patients in three different continents [9]. Some concern has arisen lately on the validity of these registries since selection bias; it has been reported that two registries included MI cases at lower risk than MI cases in the reference population [10–13]. Moreover, these registries often do not include in the performance reports the patients experiencing either a delay in hospital arrival or an inter-hospital transfer during the acute phase, making it difficult to monitor the entire process of acute care [8, 14]. These restrictions may constitute key limitations when these registries are used for public health planning [15] or for outcome research [16], calling for alternative approaches. Population-based registries provided an enormous contribution to cardiovascular diseases epidemiology. The methodological approach in these registries is to utilize consolidated administrative databases as sources of event identification and obtain clinical details via case note abstraction (‘cold-case pursuit [17]’). In an acute condition such as the myocardial infarction, the retrospective study design mitigates against selection bias since all hospitalized cases with a discharge diagnosis of MI are included [18] and it may provide longitudinal data on the patient. Challenging aspects to this approach are the completeness of data collection, since information on quality of care is retrospectively extracted from case records, as well as timing in final results availability. Recently, a population-based approach has been shown to be feasible to describe quality of cancer care [19]. In this methodological paper, we aim to evaluate the feasibility of monitoring the quality of care in hospitalized acute MI events using a population-based registry established in the Province of Varese, Northern Italy. We focus on the data completeness for critical event times and eligibility to action or treatment criteria. In addition, we characterize all patients being discharged with a diagnosis of ST-elevated MI (STEMI), including those experiencing a delay in hospital arrival or an intra-hospital transfer, to document the importance of monitoring the entire process of care. Methods Study setting and event identification The MI CAMUNI (CArdiovascular Monitoring Unit in Northern Italy) Register is a population-based registry of MI events occurred to 35–79-years-old residents in the Varese Province in 2007–2008, extending the WHO MONICA MI Register established in the nearby Brianza since 1985 [20]. The Varese Province is located in the Lombardia Region between Milan and the Swiss border. Since the clinical performance measures are referring to the in-hospital setting, for the purpose of the present report we focus on hospitalized cases. The data source for the in-hospital events was the Regional Hospital Discharge Diagnoses (HDD) database. Selection criterion was the presence of an ICD-9 410.xx (except 410.x2) as main discharge diagnosis. The HDD database is an administrative healthcare archive held mainly for hospital reimbursement, with no detailed clinical information on the patients. Being based upon a retrospective data collection, the register did not need a formal approval by an Institutional Review Board. Data collection Case reports of selected hospitalizations were reviewed by trained medical personnel to extract relevant information to assess performance: patient demographics, clinical features (symptom-onset time, first medical contact details, initial ECG, hospital arrival, out-of-hospital cardiac arrest, symptoms and cardiac markers), in-hospital management (aspirin at arrival, fibrinolysis, percutaneous coronary angioplasty (PCI), left ventricular systolic function evaluation), clinical reasons for not performing or delaying a treatment, previous history of MI and cardiovascular disease risk factors (smoking status, diabetes, hypertension and family history of MI), drug treatment at admission and at discharge. Approximate event time was collected when it was the only available information (i.e. time between symptoms onset and hospital arrival less than 12 h). The study personnel received a study manual and periodical training during the 12-month data collection period; a few meetings were dedicated to discuss controversial cases. During data collection, data quality checks were carried out to detect missing values and data inconsistency. All data violations were solved and missing values confirmed whenever information was not present in the case history. A final data quality assessment is available online [21]. Definition of acute MI events Hospitalizations due to non-acute (elective) interventions such as coronary angiography or coronary artery bypass grafting were excluded (n = 83). Consecutive hospitalizations on the same patient due to hospital transfer were considered as one acute MI event (n = 39, of which 17 within the first 24 h). Access to the Emergency Department of one hospital followed by transfer to another study hospital was considered as transferred event. Each event ends with either the patient’s death, transfer to another hospital for elective interventions or rehabilitation, or discharge to home. Definition of quality of care Quality of care was assessed according to the American Heart Association performance measures [22], which quantitatively detail the adherence to treatment guidelines during the acute phase of the event (aspirin administration at arrival, reperfusion treatment for STEMI patients), the hospital stay (LDL-cholesterol assessment, evaluation of left ventricular systolic function), and at discharge (secondary prevention: drug prescription, smoking cessation advice, patient’s referral to cardiac rehabilitation). Performance measures were expressed as either (i) proportion of patients treated according to guidelines relative to those eligible for action/treatment; or (ii) timeliness of reperfusion therapy (median time to fibrinolysis/PCI) among re-perfused STEMI patients. Each indicator requires to define the ‘eligible population’, i.e. those who should have been treated according to guidelines excluding subjects with specific contraindications to action/treatment (allergy for instance) or other non-medical reasons leading to ineligibility. The complete list of performance measures and eligibility criteria is provided as Supplementary Table 1. It is worth noting that since our approach allowed collecting data at all hospital facilities for transferred patients, there was no need to exclude these patients in calculating the quality of care measures, including aspirin treatment at arrival and door-to-needle and door-to-balloon times. Statistical analysis Data completeness was presented as the prevalence of missing data on (i) critical date/times, for transferred and not transferred events; and (ii) on eligibility criteria for each performance measure. In addition, we compared clinical and demographic characteristics of four categories of STEMI patients: late hospital arrival (>12 h of symptoms onset); arrival within 12 h but in a non-PCI capable hospital; arrival within 12 h to a PCI-capable hospital but not receiving primary PCI; arrival to a PCI-capable hospital within 12 h and receiving primary (i.e. within 24 h of hospital arrival) PCI. Categories have been defined following clinical eligibility to primary PCI treatment [22]. The order of presentation corresponds to an increasing likelihood of the patients to be represented by clinical registries, the first category being always excluded and the last satisfying inclusion to procedural registries of PCI patients. Major demographic and clinical characteristics were summarized as mean (standard deviation) or proportion. We tested the null hypothesis of no difference among the groups using an analysis of variance or a chi-square test for continuous and categorical variables, respectively. All the analyses were conducted using SAS software version 9.4. Results Characteristics of the study population From the Regional Hospital Discharge Diagnoses (HDD) database, we identified n = 1830 hospitalizations with a diagnosis of 410.xx at discharge in one of the 11 hospitals in the study area (one university hospital and four PCI referral facilities) occurred to residents in the Varese Province (Fig. 1). We were able to access 1821 of the 1830 hospital clinical records (99.5%), corresponding to 1699 acute MI events after the case history review (Fig. 1). The patients had an average age of 64.6 years (SD: 10.6 years); 72% of them were men; 17% (n = 300) were transferred from the Emergency Department of another hospital, 65% (n = 1098) had a discharge diagnosis of STEMI, and 3.7% (n = 63) died in hospital. Figure 1 View largeDownload slide Event identification from the Hospital Discharge Records database, data collection and definition of the number of acute MI events after considering in-hospital transfers at admission, including transfer from Emergency Department of another facility. The CAMUNI registry. Figure 1 View largeDownload slide Event identification from the Hospital Discharge Records database, data collection and definition of the number of acute MI events after considering in-hospital transfers at admission, including transfer from Emergency Department of another facility. The CAMUNI registry. Data completeness Table 1 reports the percentages of available data for the entire sample and for transferred and non-transferred patients. Exact time of first medical contact and of hospital admission was available in 82% and 96% of patients, respectively, with no difference according to transfer status (P-values >0.5). For transferred events it was more difficult to establish the exact timing of symptoms onset (67% vs. 74% in non-transferred patients; P = 0.02). However, information on presentation time within 12 h of symptoms onset was available for 98% of all events, regardless of transfer status. The exact time of balloon inflation was available in 77% of STEMI patients that underwent PCI, with no difference according to transfer status (P = 0.3). Nevertheless, it was possible to establish whether a PCI occurred in the first 24 h of hospital arrival in all the patients, regardless of inter-hospital transfer. Finally, exact time of fibrinolysis had the highest prevalence of missing data (66%), especially so for treatment administered in the Emergency Department before inter-hospital transfer (56%: chi-square test P-value = 0.07). Table 1 Proportion of data availability (%) for critical event times, for all acute MI events and by type of hospital admission. The CAMUNI Registry   Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0      Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0    aExact time is defined as the presence in the case history of the date/time of action/procedure; approximate time is referred to the presence in the case history of enough information to establish at least whether an action/treatment happens within a certain amount of time. bTransferred patients: arrival at the emergency department of the first hospital. cAmong those receiving PCI during the first 24 h of hospital arrival. dChi-square test P-value for testing a difference in data completeness according to transfer status. Table 1 Proportion of data availability (%) for critical event times, for all acute MI events and by type of hospital admission. The CAMUNI Registry   Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0      Acute MI events  Type of hospital admission  P-valued  Non-transferred events  Transferred from ED/Hospital  (n = 1699)  (n = 1399)  (n = 300)  Critical event times           Exact timea (%)    Date/time of symptoms onset  72.9  74.1  67.3  0.02    Date/time of first medical contact  81.5  81.8  80.0  0.5    Date/time of emergency department arrivalb  95.9  97.3  89.3  <0.0001    Date/time of hospital admission  96.4  96.4  96.0  0.8    Date/time of PCI (balloon)  77.0  76.4  79.7  0.3    Date/time of Fibrinolysis (needle)  66.1  70.9  55.6  0.1   Approximate timea (%)    Hospital arrival before 12 h of symptoms onset  98.7  98.9  97.7  0.1    Survival to the first 24 h of hospital arrival  99.4  100.0  96.7  <0.0001    PCI during the first 24 h of hospital arrivalc  100.0  100.0  100.0    aExact time is defined as the presence in the case history of the date/time of action/procedure; approximate time is referred to the presence in the case history of enough information to establish at least whether an action/treatment happens within a certain amount of time. bTransferred patients: arrival at the emergency department of the first hospital. cAmong those receiving PCI during the first 24 h of hospital arrival. dChi-square test P-value for testing a difference in data completeness according to transfer status. Table 2 details the proportion of missing data on eligibility criteria and on action/treatment for each performance measure. Missing data on the eligible criteria exceeded 10% for the ‘statin prescription at discharge’ measure, due to 23% of events with no LDL-cholesterol assessment during the hospital stay. More than 5% of missing data were observed for left ventricular systolic dysfunction (9.4%), lipid-lowering therapy as pre-arrival medication (8.3%), smoking status (7.4%) and Warfarin/Coumadin as a pre-arrival medication (7.4%; performance measure ‘Aspirin at arrival’). The proportion of eligible MI events depends on the population of interest, ranging from 18% for ACE-inhibitor prescription for patients with left ventricular systolic dysfunction, to 100% for left ventricular function evaluation during the hospital stay. Finally, among eligible, the proportion of missing data on action/treatment was below 5% for all the indicators. Table 2 Data completeness of eligibility criteria and action/treatment for each performance measure. The CAMUNI Registry Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  NA: not applicable (no patient’s based eligibility criteria). LVSD, left ventricular systolic dysfunction (documented left ventricular ejection fraction < 40%). ACE-I, angiotensin-converting-enzyme inhibitor. aThis performance measure is not relevant for in-hospital deaths during the first 24 h (n = 32). bThis performance measure is relevant for STEMI events only (n = 1098). cThis performance measure is not relevant for in-hospital deaths (n = 63). See Table 1 in the supplementary material for a complete description of patient’s based exclusion criteria for each performance measure. dPercentage of MI events with missing data on eligibility criteria, on the total number of acute MI events (column 1). ePercentage of MI events with missing data on action/treatment, on the total number of eligible acute MI events (column 3). Table 2 Data completeness of eligibility criteria and action/treatment for each performance measure. The CAMUNI Registry Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  Performance measure  # of acute MI events  Missing data on eligibility criteria (%)d  # of eligible MI events  Missing data on action/treatment (%)e  Proportion of patients receiving Aspirin during the first 24 h  1667a  124 (7.4)  1443  38 (2.7)  Proportion of STEMI patients receiving reperfusion treatment  1098b  16 (1.5)  893  0 (0)  Proportion of patients receiving left ventricular function evaluation during hospital stay  1636c  NA  1636  0 (0)  Proportion of patients with LDL-cholesterol assessment during hospital stay  1636c  136 (8.3)  1183  0 (0)  Proportion of patients with aspirin prescription at discharge  1636c  0 (0)  1482  11 (0.7)  Proportion of patients with beta-blockers prescription at discharge  1636c  0 (0)  1634  11 (0.7)  Proportion of patients with statin prescription at discharge  1636c  389 (23.8)  928  10 (1.1)  Proportion of patients with LVSD receiving ACE-I prescription at discharge  1636c  153 (9.4)  301  10 (3.4)  Proportion of smokers receiving smoking cessation advice/counseling prior to discharge  1636c  121 (7.4)  517  0 (0)  Proportion of patients referral to cardiac rehabilitation  1636c  NA  1636  7 (0.4)  NA: not applicable (no patient’s based eligibility criteria). LVSD, left ventricular systolic dysfunction (documented left ventricular ejection fraction < 40%). ACE-I, angiotensin-converting-enzyme inhibitor. aThis performance measure is not relevant for in-hospital deaths during the first 24 h (n = 32). bThis performance measure is relevant for STEMI events only (n = 1098). cThis performance measure is not relevant for in-hospital deaths (n = 63). See Table 1 in the supplementary material for a complete description of patient’s based exclusion criteria for each performance measure. dPercentage of MI events with missing data on eligibility criteria, on the total number of acute MI events (column 1). ePercentage of MI events with missing data on action/treatment, on the total number of eligible acute MI events (column 3). Comparison of STEMI patients groups Patients’ flow-chart and likelihood of receiving on-site primary PCI are presented in Fig. 2. According to guidelines [23], STEMI patients with a time from symptom onset to hospital arrival of less than 12 h and who presented to a PCI-capable facility (n = 600, 55% of all STEMI) are eligible for on-site primary PCI. Of these, 310 patients (28% of initial STEMI population) actually received treatment; the median door-to-balloon time was 78 min (interquartile range: 50, 136 min), and 58% were treated within 90 min. Patients treated with on-site primary PCI were 4 years younger, more likely to be men and smokers, and less likely to have a history of MI, diabetes and hypertension than those who did not receive any primary reperfusion therapy (Table 3, columns (3) and (4); all P-values <0.05). In-hospital mortality was almost 4-fold larger among un-treated (8.5%) than among treated patients (2.3%; P = 0.001). Compared to patients experiencing either a delay in hospital arrival (column 1) or admitted to a non-PCI capable hospital (column 2), STEMI patients receiving primary PCI were younger (P = 0.01), less likely to have diabetes (P < 0.0001) and more likely to have used a pre-hospital medical service (P < 0.0001). Patients first admitted to a non-PCI capable hospital had twice the in-hospital mortality rate (5.1%) than treated patients. Table 3 Clinical characteristics at admission and in-hospital mortality of ST-elevation MI patients according to time between symptoms onset and hospital arrival, type of hospital arrival, and reperfusion treatment. ST-elevation MI patients with information on time of symptom onset (n = 1082). The CAMUNI Registry   Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09    Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09  aExcluding those who had fibrinolysis (n = 83) and those with treatment contraindications (n = 6) bP-value testing the difference between (4) and (3). T-test and chi-square test for continuous and categorical variables, respectively cP-value testing the difference between (4), (1) and (2) [2df test]. F-test and chi-square test for continuous and categorical variables, respectively Table 3 Clinical characteristics at admission and in-hospital mortality of ST-elevation MI patients according to time between symptoms onset and hospital arrival, type of hospital arrival, and reperfusion treatment. ST-elevation MI patients with information on time of symptom onset (n = 1082). The CAMUNI Registry   Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09    Hospital arrival within 12 hours of symptoms onset      Late hospital arrival (1)  First arrival in a non-PCI-capable hospital (2)  Arrival in a PCI-capable hospital  P-valueb  P-valuec  No primary reperfusion treatmenta (3)  Primary PCI (4)  Number of events  188 (17%)  294 (27%)  201 (18%)  310 (28%)  –  –  Characteristics at admission   Mean age at admission  65.2 (9.8)  63.4 (11.1)  66.8 (9.9)  62.3 (10.8)  <0.0001  0.01   Men, %  69.7  75.5  70.7  78.1  0.06  0.1   Pre-hospital medical service, %  17.0  25.2  35.8  41.9  0.2  <0.0001   History of MI, %  15.4  17.7  31.3  17.4  0.0003  0.8   Type II diabetes, %  29.3  20.4  18.9  12.6  0.05  <0.0001   Current smokers, %  30.9  32.3  28.4  38.1  0.02  0.2   History of hypertension, %  59.0  51.0  58.7  48.7  0.03  0.07   Family history of MI, %  37.2  33.7  35.3  41.0  0.2  0.2  Outcome   In-hospital mortality, %  2.1  5.1  8.5  2.3  0.001  0.09  aExcluding those who had fibrinolysis (n = 83) and those with treatment contraindications (n = 6) bP-value testing the difference between (4) and (3). T-test and chi-square test for continuous and categorical variables, respectively cP-value testing the difference between (4), (1) and (2) [2df test]. F-test and chi-square test for continuous and categorical variables, respectively Figure 2 View largeDownload slide Patients’ flow-chart and likelihood of receiving reperfusion treatment, and median door-to-balloon time (interquartile range) for STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset. The CAMUNI registry. Patients’ selection was made according to reference [22] for door-to-balloon time in patients receiving on-site PCI within the first 24 h of hospital arrival (primary PCI). ‘Door’ time is the arrival at the emergency department. *n = 16 MI events with no information on time of symptoms onset. Figure 2 View largeDownload slide Patients’ flow-chart and likelihood of receiving reperfusion treatment, and median door-to-balloon time (interquartile range) for STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset. The CAMUNI registry. Patients’ selection was made according to reference [22] for door-to-balloon time in patients receiving on-site PCI within the first 24 h of hospital arrival (primary PCI). ‘Door’ time is the arrival at the emergency department. *n = 16 MI events with no information on time of symptoms onset. Discussion In this paper, we document the feasibility and the importance of a population-based registry for monitoring the quality of care in hospitalized acute myocardial infarction patients, as defined by standard performance indicators [22]. Event selection was based on the main discharge diagnosis code 410.xx, which was shown to provide reliable data on acute myocardial infarction in Northern Italy; [24] we were able to review the case histories of 99% of selected cases. Data completeness in hospital-based registries is an issue whenever missing data are related to event severity [11]. In our population-based registry, we were able to establish most of the critical date/time, including presentation time and primary PCI execution, in virtually all the events and regardless of initial transfer status. One PCI-capable hospital began to report exact PCI time only from 2008 onwards, contributing to the 23% of missing date/time of PCI intervention; however missing data was independent of the patient’s characteristics including age, gender, presentation delay from symptom onset, prevalence of typical symptoms and in-hospital mortality (data not shown). Eligibility criteria are essential to carefully define the subset of patients who should have been treated according to guidelines and thus ultimately provide unbiased estimates of performance [22]. We were able to collect information on most of the eligibility criteria, as documented in the Supplementary Table 1. Although it is difficult to compare the prevalence of eligible patients in different contexts, the prevalence of smokers, aspirin allergy and left ventricular systolic dysfunction in our study was similar to other reports in Italy [25] and the USA [14]. Missing data on eligible criteria were mostly due to either lack of information on pre-arrival medication (for ASA at arrival and LDL-cholesterol assessment performance measures) or no risk evaluation during the hospital stay (no LDL-cholesterol assessment, no left ventricular systolic function evaluation). Patients may not receive an LDL-assessment because they are considered at high risk [26] and thus fulfill the treatment indication (83% prevalence of statin prescription at discharge in our registry in this group, compared to 90% in the group with LDL-cholesterol >100 mg/dl and 77% in patients with LDL-cholesterol <100 mg/dl). Implications Quality of care assessment as defined by core hospital performance measures is undertaken mostly within a hospital-based perspective which often excludes the reporting of patients transferred during the acute event phase, as well as of patients with a late hospital presentation [8, 14, 27]. In our registry, the median door-to-balloon time for STEMI patients admitted to a PCI-capable facility within 12 h of symptoms onset was 78 min, similar to what reported by procedural registries in the same period in Italy [25]. These patients were comparable to the average patient profile in the PCI procedures registry active in the Lombardia region during the same time span: mean age 63 years, 77% men, 15% with diabetes and 40% admission through the Emergency Medical System [28]. However, as documented by Fig. 2 and Table 3, they are the expression of a care pathway which selects only 28% of initial STEMI patients: any further improvement in door-to-balloon time will affect less than 1 patient out of 3. About 44% of STEMI patients, in particular older women with comorbidities and patients less likely to access to the pre-hospital medical service, experienced either a delay in hospital presentation or a first arrival to a non-PCI capable hospital. A further 18% of STEMI patients arrived to a PCI-capable hospital within 12 h of symptoms onset, but did not receive primary PCI. These were on average 4 years older; were more likely to have comorbidities (history of MI, diabetes, hypertension); and experienced an in-hospital mortality rate almost four times higher than treated patients. By design, clinical and procedural registries are not feasible to quantify nor to appropriately monitor the patients’ selection which affects the care process: in a systematic review of 129 published papers from 77 clinical registries, only 3% were aware of the possibility of selection bias [12]. The advantage of the retrospective approach is to allow a comprehensive evaluation of the entire acute care pathway, providing additional information that stakeholders and decision makers can use in an integrated disease management plan [29, 30]. The proportion of STEMI patients who arrive in a PCI-capable hospital within 12 h of symptoms onset can be an exemplification. First, the retrospective approach can be used to identify population subgroups (the elderly, women) to be targeted by specific interventions to increase symptoms awareness and facilitate the use of pre-hospital medical service. Then, the process indicator can be monitored by extending the registry over time. Finally, through record linkage to external data such as the routine mortality database or the hospital discharge records, it is possible to assess the efficacy of the intervention in the reduction of re-hospitalizations and mortality. Current limitations and future improvements in study design The study design might be improved by considering a weighted sampling of cases to reduce duration and costs of data collection, by improving the timeliness of findings availability. Data completeness depends on case histories, and might vary in other settings. Based on our previous experience on epidemiological registries, we fixed an upper age limit at 79 years due to the fact that collecting reliable data on older patients is very challenging within our system. Moreover, we limited our registry to residents in the Varese Province only (i.e. non-residents cases were excluded). Residency is a standard inclusion criterion for many epidemiologic registries. Based on our data, non-residents cases were on average younger than residents (63.6 vs. 64.9 years old at hospital admission) and many performance measures are associated with the patient’s age [31, 32]. Therefore this choice is not recommended in future studies but it is not likely to have affected our study implications due to the study design. Conclusions In conclusion, this study documented the feasibility and importance of a population-based registry to assess the quality of hospital management in acute MI patients. Complementary to clinical registries, the population-based approach allows monitoring of the entire acute care process, providing incremental valuable information for public health planning and resources allocation strategies. Supplementary material Supplementary material is available at International Journal for Quality in Health Care online. Acknowledgments The authors are grateful to Dr. A. Borsani and Dr. M. Bonzini for developing and testing the case report form, as well as for coordinating the data collection activities. We thank Drs. S. Mombelli, M. Casà N. Facchinetti, S. Landone, M. Conti, R. Corrao and S. Colombo for review of case histories, data collection and data checking; and Dr. M. de Biasi for data quality assessment. We also thank Dr. M. Marzegalli, Cardiology Unit of the San Carlo Hospital, Milan, and Dr. O. Leoni, Agenzia di Tutela della Salute dell’Insubria, for valuable comments on earlier drafts of this manuscript. Funding This work was supported by the Italian Ministry of Health [Grant Ricerca Finalizzata Programma Strategico-2007-2-634753]. G.V. work was supported by a grant from the Health Administration of Lombardia Region [Grant 10800/2009] as part of the ‘Osservatorio Regionale Lombardo per le malattie cardiovascolari’. Conflict of interest statement None. References 1 Lozano R, Naghavi M, Foreman K et al.  . Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet  2012; 380: 2095– 128. Google Scholar CrossRef Search ADS PubMed  2 Writing Group Members. Mozaffarian D, Benjamin EJ, Go AS et al.  . American Heart Association Statistics Committee; Stroke Statistics Subcommittee. Heart Disease and Stroke Statistics-2016 update: a report from the American Heart Association. Circulation  2016; 133: e38– 360. Google Scholar CrossRef Search ADS PubMed  3 Gitt AK, Bueno H, Danchin N et al.  . The role of cardiac registries in evidence-based medicine. Eur Hearth J  2010; 31: 525– 29. Google Scholar CrossRef Search ADS   4 Gibson CM, Pride YB, Frederick PD et al.  . Trends in reperfusion strategies, door-to-needle and door-to-balloon times, and in-hospital mortality among patients with ST-segment elevation myocardial infarction enrolled in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J  2008; 156: 1035– 44. Google Scholar CrossRef Search ADS PubMed  5 Laskey W, Spence N, Zhao X et al.  . Regional differences in quality of care and outcomes for the treatment of acute coronary syndromes. an analysis from the get with the Guidelines Coronary Artery Disease Program. Crit Pathways Cardiol  2010; 9: 1– 7. Google Scholar CrossRef Search ADS   6 Vogeli C, Kang R, Landrum MB et al.  . Quality of care provided to individual patients in US hospitals. Results from an analysis of National Hospital Quality Alliance Data. Med Care  2009; 47: 591– 9. Google Scholar CrossRef Search ADS PubMed  7 Widimsky P, Wijns W, Fajadet J et al.  . European Association for Percutaneous Cardiovascular Interventions. Reperfusion therapy for ST elevation acute myocardial infarction in Europe: description of the current situation in 30 countries. Eur Heart J  2010; 31: 943– 57. Google Scholar CrossRef Search ADS PubMed  8 Balzi D, Di Bari M, Barchielli A et al.  . Should we improve the management of NSTEMI? Results from the population-based ‘acute myocardial infarction in Florence 2’ (AMI-Florence 2) registry. Intern Emerg Med  2013; 8: 725– 33. Google Scholar CrossRef Search ADS PubMed  9 GRACE investigators. Rationale and design of the GRACE (Global Registry of Acute Coronary Events) Project: a multinational registry of patients hospitalized with acute coronary syndromes. Am Heart J  2001; 141: 190– 9. CrossRef Search ADS PubMed  10 Krumholz HM. Registries and selection bias: the need for accountability. Circ Cardiovasc Qual Outcomes  2009; 2: 517– 8. Google Scholar CrossRef Search ADS PubMed  11 McNamara RL. Cardiovascular registry research comes of age. Heart  2010; 96: 908– 10. Google Scholar CrossRef Search ADS PubMed  12 Ferreira-Gonzalez I, Marsal JR, Mitjavila F et al.  . Patient registries of acute coronary syndrome. Assessing or biasing the clinical real world data? Circ Cardiovasc Qual Outcomes  2009; 2: 540– 7. Google Scholar CrossRef Search ADS PubMed  13 Rosvall M, Ohlsson H, Hansen O et al.  . Auditing patient registration in the Swedish quality register for acute coronary syndrome. Scand J Public Health  2010; 38: 533– 40. Google Scholar CrossRef Search ADS PubMed  14 Bernheim S, Wang Y, Bradley EH et al.  . Who is missing from the measure? Trends in proportion and treatment of patients potentially excluded from publicly-reported quality measures. Am Heart J  2012; 160: 943– 50. Google Scholar CrossRef Search ADS   15 Larsson S, Lawyer P for the Boston Consulting Group. Improving health care value. The case for disease registries. [https://www.bcgperspectives.com/content/articles/health_care_payers_providers_biopharma_improving_health_care_value_disease_registries/] (December 11, 2017, date last accessed). 16 Roger VL. Outcome research and epidemiology. The synergy between public health and clinical practice. Circ Cardiovasc Qual Outcomes  2011; 4: 257– 9. Google Scholar CrossRef Search ADS PubMed  17 Tunstall-Pedoe H. Problems with criteria and quality control in the registration of coronary events in the MONICA study. Acta Med Scand Suppl  1988; 728: 17– 25. Google Scholar PubMed  18 Brieger D, Aliprandi-Costa B. Developments in procedural and disease registries: a focus on coronary artery disease. Curr Opin Cardiol  2013; 28: 405– 10. Google Scholar CrossRef Search ADS PubMed  19 Caldarella A, Amunni G, Angiolini C et al.  . Feasibility of evaluating quality cancer care using registry data and electronic health records: a population-based study. Int J Qual Health Care  2012; 24: 411– 8. Google Scholar CrossRef Search ADS PubMed  20 Veronesi G, Ferrario MM, Chambless LE et al.  . The effect of revascularization procedures on myocardial infarction incidence rates and time trends: The MONICA-Brianza and CAMUNI MI registries in Northern Italy. Ann Epidemiol  2012; 22: 547– 53. Google Scholar CrossRef Search ADS PubMed  21 Veronesi G, De Biasi M, Ferrario MM Registry of hospitalized Acute Myocardial infarction in the Varese Province: Data Quality Assessment. Available at: http://www4.uninsubria.it/on-line/home/naviga-per-tema/ricerca-scientifica/centri-di-ricerca/centro-di-ricerca-in-epidemiologia-e-medicina-preventiva-epimed/documento307171.html (December 11, 2017, date last accessed). 22 Writing Committee to Develop Performance Measures for ST-Elevation and Cardiology/American Heart Association Task Force on Performance Measures. ACC/AHA 2008 performance measures for adults with ST-Elevation and Non ST-Elevation Myocardial Infarction. A report of the American College of Cardiology/American Heart Association Task Force on Performance Measures. Circulation  2008; 118: 2596– 648. CrossRef Search ADS PubMed  23 Piepoli MF, Hoes AW, Agewall S et al.  . 2016 European Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J  2016; 37: 2315– 81. Google Scholar CrossRef Search ADS PubMed  24 Brocco S, Fedeli U, Schievano E et al.  . Effect of the new diagnostic criteria for ST-elevation and non-ST elevation acute myocardial infarction on 4-year hospitalization: an analysis of hospital discharge records in the Veneto Region. J Cardiovasc Med  2006; 7: 45– 50. Google Scholar CrossRef Search ADS   25 Manari A, Ortolani P, Guastaroba P et al.  . Clincal impact of an inter-hospital transfer strategy in patients with ST-elevation myocardial infarction undergoing primary angioplasty: the Emilia-Romagna ST-segment elevation acute myocardial infarction network. Eur Hearth J  2008; 29: 1834– 42. Google Scholar CrossRef Search ADS   26 Javed U, Deedwania PC, Bhatt DL et al.  . Use of intensive lipid-lowering therapy in patients hospitalized with acute coronary syndrome: an analysis of 65 396 hospitalizations from 334 hospitals participating in Get With The Guidelines (GWTG). Am Heart J  2011; 161: 418– 24. Google Scholar CrossRef Search ADS PubMed  27 Fonarow GC, Gregory T, Driskill M et al.  . Hospital certification for optimizing cardiovascular disease and stroke quality of care and outcomes. Circulation  2010; 122: 2459– 69. Google Scholar CrossRef Search ADS PubMed  28 Martinoni A, De Servi S, Boschetti E et al.  . Lombardima Study Group. Importance and limits of pre-hospital electrocardiogram in patients with ST elevation myocardial infarction undergoing percutaneous coronary angioplasty. Eur J Cardiovasc Prev Rehabil  2011; 18: 526– 32. Google Scholar CrossRef Search ADS PubMed  29 Romeyke T, Stummer H. Clinical pathways as instruments for risk and cost management in hospitals - a discussion paper. Glob J Health Sci  2012; 4: 50– 9. Google Scholar CrossRef Search ADS PubMed  30 Bufalino VJ, Masoudi FA, Stranne SK et al.  . for the American Heart Association Advocacy Coordinating Committee. The American Heart Association’s recommendations for expanding the applications of existing and future clinical registries: a policy statement from the American Heart Association. Circulation  2011; 123: 2167– 79. Google Scholar CrossRef Search ADS PubMed  31 Peterson ED, Shah BR, Parsons L et al.  . Trends in quality of care for patients with acute myocardial infarction in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J  2008; 156: 1045– 55. Google Scholar CrossRef Search ADS PubMed  32 Schoenenberger AW, Radovanovic D, Stauffer JC et al.  . Acute Myocardial Infarction in Switzerland Plus Investigators. Age-related differences in the use of guideline-recommended medical and interventional therapies for acute coronary syndromes: a cohort study. J Am Geriatr Soc  2008; 56: 510– 6. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

International Journal for Quality in Health CareOxford University Press

Published: Feb 21, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off