FULL-LENGTH ORIGINAL RESEARCH
Prevalence and incidence of epilepsy in a well-defined population of Northern Italy Giorgia Giussani, Carlotta Franchi, Paolo Messina, Alessandro Nobili, Ettore Beghi, and 1the EPIRES Group Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
SUMMARY
Giorgia Giussani is an investigator at IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Italy.
Objective: To calculate prevalence and incidence of epilepsy using administrative records. Methods: Claim records from the administrative district of Lecco, Northern Italy (population 311,637; 2001 census), collected during the years 2000–2008, were the data source. Patients of all ages were included. Based on previous findings from our group, the most accurate algorithm to detect epilepsy was the combination of electroencephalography (EEG) (ad hoc code) (at least one during the study period) and antiepileptic drugs (AEDs) (ATC code) (taken in 2008). Using this algorithm, the prevalence of epilepsy for the year 2008 was calculated. The reference population for prevalence was the population residing in the study area during the year 2008. Incident epilepsy cases were a subset of prevalent cases among patients not traced in the years 2000 through 2003. Average annual incidence rates were calculated for 2004 through 2008, taking for reference the person-years of exposure in the resident population. We calculated crude, adjusted (using positive and negative predictive values), and standardized (to the Italian and World population) prevalence and incidence. Results: In 2008, 1,504 patients met the inclusion criteria, giving a prevalence of 4.57 per 1,000 (women 4.26; men 4.89). Prevalence tended to rise slightly with age. There were 864 incident cases, giving an average annual incidence of 53.41 per 100,000 (women 50.98; men 55.95). Incidence rates peaked in the elderly. The adjusted prevalence was 4.42 and the adjusted incidence 47.05. Standardized prevalence and incidence were, respectively, 4.30 per 1,000 and 48.35 per 100,000 (Italian population) and 3.79 per 1,000 and 44.74 per 100,000 (World population). Significance: The prevalence of epilepsy in the Lecco district was comparable to other studies, whereas the incidence was among the highest. With adjustments, administrative records are a cost-effective instrument to monitor epilepsy frequency. KEY WORDS: Epilepsy, Prevalence, Incidence, Administrative data, Validity.
Accepted July 9, 2014; Early View publication August 4, 2014. IRCCS Mario Negri Institute for Pharmacological Research, Milan, Italy 1 EPIRES Group: Elio Agostoni, Larissa Airoldi, Francesco Basso, Valentina Canelli, Marialuisa Carpanelli, Mariolina Di Stefano, Andrea Magnoni, Ottaviano Martinelli, Mariagrazia Montesano, Andrea Rigamonti, Silvia Romi, Andrea Salmaggi, Lorenzo Stanzani, Cristina Volpe, Nicoletta Zanotta, Claudio Zucca. Address correspondence to Ettore Beghi, Laboratory of Neurological Disorders, IRCCS-Mario Negri Institute for Pharmacological Research, Via Giuseppe La Masa 19, 20156 Milan, Italy. E-mail: [email protected] Wiley Periodicals, Inc. © 2014 International League Against Epilepsy
Epilepsy is a chronic clinical condition with significant impact on the patients and the health care systems. Epidemiologic surveys are useful for defining the burden of epilepsy in a population. This is particularly helpful for aging populations, as several epileptogenic conditions are age-related. The focus of epidemiologic surveys is to identify virtually all patients from representative population samples to give correct estimates of the prevalence and incidence of the disease. Case ascertainment methods include direct interviews and retrospective reviews of diagnostic codes, medical notes, antiepileptic drugs (AEDs), and electroencephalography (EEG) prescriptions, in variable combinations.
1526
1527 Prevalence and Incidence of Epilepsy However, investigation of multiple sources is time-consuming, frequently expensive, and depends on the collaboration of doctors, patients, and health care administrators. Published reports on the prevalence and incidence of epilepsy in Italy are therefore few and, with one exception,1 are based on small samples2,3 or selected subgroups of the general population.4 Administrative data based on the health care resources consumed for the diagnosis and treatment of the disease are a valuable source of epidemiologic findings because they are obtained from large representative population samples, are routinely collected, and can be obtained at almost no cost. However, before they can be used as disease tracers, administrative data need to be validated to test the accuracy of the information. In Italy, health care is provided almost free of charge by the National Health Service (NHS) to the entire population. In addition, patients with chronic diseases are entitled to full exemption from paying medical consultations, diagnostic tests, and drugs when indicated for the management of the disease. Administrative data are collected in the health facilities to be refunded for care to Italian citizens. This system was activated in the year 1975 and has been fully operative since 2000. Thus, it can be assumed that almost patients diagnosed with epilepsy can be traced through the public health care system. The main aim of this study was to use information from the administrative records of a well-defined geographic area in Northern Italy to calculate the prevalence and incidence of epilepsy. Specific aims included calculating age-specific and sex-specific prevalence and incidence, with special attention to the elderly.
Materials and Methods Study area and population As a data source we used the records of the claims for health care services in the regional database from the administrative district of Lecco, Lombardy Region, Northern Italy (population 311,637; 2001 census), collected during the years 2000 through 2008. The database includes prescription records containing a personal identifier code, sex, date of birth, International Classification of Diseases, 9th edition (ICD-9-CM) codes for the discharge diagnoses of all hospital admissions, prescriptions for diagnostic tests (including EEG), disease-specific exemption codes (i.e., codes consenting free access to laboratory, electrophysiologic and imaging tests, consultations, and drugs for epilepsy), and drugs dispensed. Only drugs provided free of charge by the Italian NHS, which include all marketed AEDs, are collected. All drugs prescribed are classified according to the international Anatomic Therapeutic Classification system (ATC).5 The structure of the local population in the year 2008 is illustrated in Table 1. The study area is in Northern Italy,
covering a surface of about 816 km2 and is mostly hilly (Fig. 1). Industrial and agricultural activities are both present. The local population is almost entirely of Caucasian origin (96%) and is fairly stable, with a migration rate of 3.28% for the year 2008 (emigration, 1.20%; immigration, 2.08%). Diagnosis of epilepsy and seizures In keeping with the requirements for epidemiologic studies, epilepsy was defined as a condition marked by two or more unprovoked epileptic seizures 24+ h apart with neurologic confirmation.6 Isolated seizures and acute symptomatic seizures were classified separately and excluded. These included single unprovoked epileptic seizures and, respectively, seizures occurring in close temporal association with an acute systemic, metabolic, or toxic insult or an acute central nervous system (CNS) insult (infection, stroke, cranial trauma, intracerebral hemorrhage).6 The validity of the diagnosis was tested by direct examination of the general practitioner’s (GP’s) medical records. In Italy the diagnosis of epilepsy is not established by the GP unless there is a confirmation by a neurologist (or child neurologist). Validation of the administrative records An unselected sample of the data source was validated in a previous study.7 The algorithm to detect epilepsy was chosen as the most predictive combination of the following codes: (1) ICD-9-CM code for epilepsy (345.x) or 333.2 (myoclonus) or 780.3 (convulsions, febrile or afebrile) or 779.0 (neonatal seizures) or 781.0 (spasms, other abnormal involuntary movements); these codes are only present for hospitalized patients; (2) exemption code for epilepsy; (3) having at least one EEG (ad hoc code); ad (4) taking one or more AEDs (ATC codes) as monotherapy or in variable combinations. Age, sex, and the patient’s GPs were also recorded. Cases (71) were ascertained from 15,728 affiliates of 11 GPs (including three pediatricians) working in the study area. The most valid and conservative algorithm to detect patients with epilepsy was the combination of EEG and AEDs. An algorithm combining EEGs with antiepileptic drugs over a 9-year period was a plausible diagnostic surrogate for several reasons. First of all, in Italy, the GPs don’t confirm the diagnosis of epilepsy without consulting a neurologist (or child neurologist). Second, a patient with epilepsy is generally followed by the (child) neurologist, who decides if and when to perform new diagnostic tests (including EEG studies) and treatment changes. Third, EEG testing is performed at the time of the diagnosis and during followup when there are diagnostic or seizure control problems. Fourth, in patients with prolonged seizure remission, an EEG is generally obtained to predict seizure relapse before deciding drug discontinuation. For these reasons, we expect that EEG testing was performed at least once in the 9-year period in virtually all patients with epilepsy. For Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
789 715 1504
61 174 265 186 103
49 151 243 166 106
110 325 508 352 209
161,385 168,001 329,386
24,084 38,618 51,599 36,810 10,274
23,111 36,552 49,578 39,377 19,383
47,195 75,170 101,177 76,187 29,657
y y y
y y y
y y y
n
2.33 4.32 4.02 4.62 7.05
2.12 4.13 4.90 4.22 5.47
2.53 4.51 5.14 5.05 10.03
4.89 4.26 4.57
Pr*1,000 (crude)
193–2.81 3.88–4.82 4.60–5.48 4.16–5.13 6.16–8.07
1.60–2.80 3.52–4.84 4.32–5.56 3.62–4.91 4.52–6.61
1.97–3.25 3.89–5.22 4.55–5.79 4.38–5.83 8.27–12.14
4.56–5.24 3.96–4.58 4.34–4.80
95% CIs
1.000 0.750 0.714 0.700 0.571
1.000 1.000 0.750 0.818 0.750
1.000 0.600 0.667 0.556 0.333
0.607 0.824 0.726
PPV
1.000 0.999 0.999 0.999 0.995
0.999 0.999 0.999 0.999 0.996
1.000 0.999 1.000 0.998 0.993
0.999 0.999 0.999
NPV
110.00 243.75 362.86 246.40 119.43
49.00 151.00 182.25 135.82 79.50
61.00 104.40 176.67 103.33 34.33
479.04 588.82 1091.61
Expected true positives
19.07 47.90 69.68 96.82 135.81
18.86 24.60 46.28 24.42 68.68
0.00 23.37 23.31 71.95 66.48
181.49 183.76 365.28
Expected false negatives
129.07 291.65 432.54 343.22 255.24
67.86 175.60 228.53 160.23 148.18
61.00 127.77 199.98 175.29 100.81
660.53 772.58 1456.89
Expected patients with epilepsy
2.73 3.88 4.28 4.50 8.61
2.94 4.80 4.61 4.07 7.64
2.53 3.31 3.88 4.76 9.81
4.09 4.60 4.42
Prevalence (adjusted)
N, Number of persons in the 2008 population; n, no. of cases; Pr*1,000, prevalence per 1,000; PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; y, years.
Men Women Total Men
Prevalence and incidence of epilepsy in a well-defined population of Northern Italy Giorgia Giussani, Carlotta Franchi, Paolo Messina, Alessandro Nobili, Ettore Beghi, and 1the EPIRES Group Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
SUMMARY
Giorgia Giussani is an investigator at IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Italy.
Objective: To calculate prevalence and incidence of epilepsy using administrative records. Methods: Claim records from the administrative district of Lecco, Northern Italy (population 311,637; 2001 census), collected during the years 2000–2008, were the data source. Patients of all ages were included. Based on previous findings from our group, the most accurate algorithm to detect epilepsy was the combination of electroencephalography (EEG) (ad hoc code) (at least one during the study period) and antiepileptic drugs (AEDs) (ATC code) (taken in 2008). Using this algorithm, the prevalence of epilepsy for the year 2008 was calculated. The reference population for prevalence was the population residing in the study area during the year 2008. Incident epilepsy cases were a subset of prevalent cases among patients not traced in the years 2000 through 2003. Average annual incidence rates were calculated for 2004 through 2008, taking for reference the person-years of exposure in the resident population. We calculated crude, adjusted (using positive and negative predictive values), and standardized (to the Italian and World population) prevalence and incidence. Results: In 2008, 1,504 patients met the inclusion criteria, giving a prevalence of 4.57 per 1,000 (women 4.26; men 4.89). Prevalence tended to rise slightly with age. There were 864 incident cases, giving an average annual incidence of 53.41 per 100,000 (women 50.98; men 55.95). Incidence rates peaked in the elderly. The adjusted prevalence was 4.42 and the adjusted incidence 47.05. Standardized prevalence and incidence were, respectively, 4.30 per 1,000 and 48.35 per 100,000 (Italian population) and 3.79 per 1,000 and 44.74 per 100,000 (World population). Significance: The prevalence of epilepsy in the Lecco district was comparable to other studies, whereas the incidence was among the highest. With adjustments, administrative records are a cost-effective instrument to monitor epilepsy frequency. KEY WORDS: Epilepsy, Prevalence, Incidence, Administrative data, Validity.
Accepted July 9, 2014; Early View publication August 4, 2014. IRCCS Mario Negri Institute for Pharmacological Research, Milan, Italy 1 EPIRES Group: Elio Agostoni, Larissa Airoldi, Francesco Basso, Valentina Canelli, Marialuisa Carpanelli, Mariolina Di Stefano, Andrea Magnoni, Ottaviano Martinelli, Mariagrazia Montesano, Andrea Rigamonti, Silvia Romi, Andrea Salmaggi, Lorenzo Stanzani, Cristina Volpe, Nicoletta Zanotta, Claudio Zucca. Address correspondence to Ettore Beghi, Laboratory of Neurological Disorders, IRCCS-Mario Negri Institute for Pharmacological Research, Via Giuseppe La Masa 19, 20156 Milan, Italy. E-mail: [email protected] Wiley Periodicals, Inc. © 2014 International League Against Epilepsy
Epilepsy is a chronic clinical condition with significant impact on the patients and the health care systems. Epidemiologic surveys are useful for defining the burden of epilepsy in a population. This is particularly helpful for aging populations, as several epileptogenic conditions are age-related. The focus of epidemiologic surveys is to identify virtually all patients from representative population samples to give correct estimates of the prevalence and incidence of the disease. Case ascertainment methods include direct interviews and retrospective reviews of diagnostic codes, medical notes, antiepileptic drugs (AEDs), and electroencephalography (EEG) prescriptions, in variable combinations.
1526
1527 Prevalence and Incidence of Epilepsy However, investigation of multiple sources is time-consuming, frequently expensive, and depends on the collaboration of doctors, patients, and health care administrators. Published reports on the prevalence and incidence of epilepsy in Italy are therefore few and, with one exception,1 are based on small samples2,3 or selected subgroups of the general population.4 Administrative data based on the health care resources consumed for the diagnosis and treatment of the disease are a valuable source of epidemiologic findings because they are obtained from large representative population samples, are routinely collected, and can be obtained at almost no cost. However, before they can be used as disease tracers, administrative data need to be validated to test the accuracy of the information. In Italy, health care is provided almost free of charge by the National Health Service (NHS) to the entire population. In addition, patients with chronic diseases are entitled to full exemption from paying medical consultations, diagnostic tests, and drugs when indicated for the management of the disease. Administrative data are collected in the health facilities to be refunded for care to Italian citizens. This system was activated in the year 1975 and has been fully operative since 2000. Thus, it can be assumed that almost patients diagnosed with epilepsy can be traced through the public health care system. The main aim of this study was to use information from the administrative records of a well-defined geographic area in Northern Italy to calculate the prevalence and incidence of epilepsy. Specific aims included calculating age-specific and sex-specific prevalence and incidence, with special attention to the elderly.
Materials and Methods Study area and population As a data source we used the records of the claims for health care services in the regional database from the administrative district of Lecco, Lombardy Region, Northern Italy (population 311,637; 2001 census), collected during the years 2000 through 2008. The database includes prescription records containing a personal identifier code, sex, date of birth, International Classification of Diseases, 9th edition (ICD-9-CM) codes for the discharge diagnoses of all hospital admissions, prescriptions for diagnostic tests (including EEG), disease-specific exemption codes (i.e., codes consenting free access to laboratory, electrophysiologic and imaging tests, consultations, and drugs for epilepsy), and drugs dispensed. Only drugs provided free of charge by the Italian NHS, which include all marketed AEDs, are collected. All drugs prescribed are classified according to the international Anatomic Therapeutic Classification system (ATC).5 The structure of the local population in the year 2008 is illustrated in Table 1. The study area is in Northern Italy,
covering a surface of about 816 km2 and is mostly hilly (Fig. 1). Industrial and agricultural activities are both present. The local population is almost entirely of Caucasian origin (96%) and is fairly stable, with a migration rate of 3.28% for the year 2008 (emigration, 1.20%; immigration, 2.08%). Diagnosis of epilepsy and seizures In keeping with the requirements for epidemiologic studies, epilepsy was defined as a condition marked by two or more unprovoked epileptic seizures 24+ h apart with neurologic confirmation.6 Isolated seizures and acute symptomatic seizures were classified separately and excluded. These included single unprovoked epileptic seizures and, respectively, seizures occurring in close temporal association with an acute systemic, metabolic, or toxic insult or an acute central nervous system (CNS) insult (infection, stroke, cranial trauma, intracerebral hemorrhage).6 The validity of the diagnosis was tested by direct examination of the general practitioner’s (GP’s) medical records. In Italy the diagnosis of epilepsy is not established by the GP unless there is a confirmation by a neurologist (or child neurologist). Validation of the administrative records An unselected sample of the data source was validated in a previous study.7 The algorithm to detect epilepsy was chosen as the most predictive combination of the following codes: (1) ICD-9-CM code for epilepsy (345.x) or 333.2 (myoclonus) or 780.3 (convulsions, febrile or afebrile) or 779.0 (neonatal seizures) or 781.0 (spasms, other abnormal involuntary movements); these codes are only present for hospitalized patients; (2) exemption code for epilepsy; (3) having at least one EEG (ad hoc code); ad (4) taking one or more AEDs (ATC codes) as monotherapy or in variable combinations. Age, sex, and the patient’s GPs were also recorded. Cases (71) were ascertained from 15,728 affiliates of 11 GPs (including three pediatricians) working in the study area. The most valid and conservative algorithm to detect patients with epilepsy was the combination of EEG and AEDs. An algorithm combining EEGs with antiepileptic drugs over a 9-year period was a plausible diagnostic surrogate for several reasons. First of all, in Italy, the GPs don’t confirm the diagnosis of epilepsy without consulting a neurologist (or child neurologist). Second, a patient with epilepsy is generally followed by the (child) neurologist, who decides if and when to perform new diagnostic tests (including EEG studies) and treatment changes. Third, EEG testing is performed at the time of the diagnosis and during followup when there are diagnostic or seizure control problems. Fourth, in patients with prolonged seizure remission, an EEG is generally obtained to predict seizure relapse before deciding drug discontinuation. For these reasons, we expect that EEG testing was performed at least once in the 9-year period in virtually all patients with epilepsy. For Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
Epilepsia, 55(10):1526–1533, 2014 doi: 10.1111/epi.12748
789 715 1504
61 174 265 186 103
49 151 243 166 106
110 325 508 352 209
161,385 168,001 329,386
24,084 38,618 51,599 36,810 10,274
23,111 36,552 49,578 39,377 19,383
47,195 75,170 101,177 76,187 29,657
y y y
y y y
y y y
n
2.33 4.32 4.02 4.62 7.05
2.12 4.13 4.90 4.22 5.47
2.53 4.51 5.14 5.05 10.03
4.89 4.26 4.57
Pr*1,000 (crude)
193–2.81 3.88–4.82 4.60–5.48 4.16–5.13 6.16–8.07
1.60–2.80 3.52–4.84 4.32–5.56 3.62–4.91 4.52–6.61
1.97–3.25 3.89–5.22 4.55–5.79 4.38–5.83 8.27–12.14
4.56–5.24 3.96–4.58 4.34–4.80
95% CIs
1.000 0.750 0.714 0.700 0.571
1.000 1.000 0.750 0.818 0.750
1.000 0.600 0.667 0.556 0.333
0.607 0.824 0.726
PPV
1.000 0.999 0.999 0.999 0.995
0.999 0.999 0.999 0.999 0.996
1.000 0.999 1.000 0.998 0.993
0.999 0.999 0.999
NPV
110.00 243.75 362.86 246.40 119.43
49.00 151.00 182.25 135.82 79.50
61.00 104.40 176.67 103.33 34.33
479.04 588.82 1091.61
Expected true positives
19.07 47.90 69.68 96.82 135.81
18.86 24.60 46.28 24.42 68.68
0.00 23.37 23.31 71.95 66.48
181.49 183.76 365.28
Expected false negatives
129.07 291.65 432.54 343.22 255.24
67.86 175.60 228.53 160.23 148.18
61.00 127.77 199.98 175.29 100.81
660.53 772.58 1456.89
Expected patients with epilepsy
2.73 3.88 4.28 4.50 8.61
2.94 4.80 4.61 4.07 7.64
2.53 3.31 3.88 4.76 9.81
4.09 4.60 4.42
Prevalence (adjusted)
N, Number of persons in the 2008 population; n, no. of cases; Pr*1,000, prevalence per 1,000; PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; y, years.
Men Women Total Men
