Canadian Psychiatric Association
Association des psychiatres du Canada
Perspective
On Adjusting for Life’s Confounding: Harnessing Big Data to Answer Big Problems
The Canadian Journal of Psychiatry / La Revue Canadienne de Psychiatrie 2017, Vol. 62(3) 182-185 ª The Author(s) 2016 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0706743716660710 TheCJP.ca | LaRCP.ca
Steve Kisely, MD, PhD1,2
Keywords epidemiology, surveillance, mental health services, health services research, comorbidity
A wide range of data sources are available to study the epidemiology of mental illness. Community surveys such as Alex Leighton’s Stirling County Study are the gold standard for such data sources, particularly because community surveys cover everyone, not just people seeking treatment. However, as surveys require a lot of resources, other methods are also used, each of which has strengths and weaknesses. For example, medical records contain detailed information, but this can be difficult to extract and the quality may vary. Another source entails administrative data, typically hospital separations, physician billings, ambulatory care visits, and drug databases. While such data require careful analysis with the use of multivariate or propensity score techniques to adjust for potential confounding variables, these data can be invaluable in the study of diseases with multifactorial aetiologies. This paper addresses how these large administrative databases, or big data, can help address mental health problems in situations where conventional clinical studies are of little help. Patients with chronic disease are an example, given they are often excluded from clinical trials on account of their multiple medical and psychiatric comorbidities. This is an important omission, as these diseases are major sources of disability and burden worldwide. While the causes of chronic illnesses are not fully known, a range of apparently interrelated personal, social, economic, and environmental factors are associated with their development. As much as 40% of chronic illness may be attributable to a small number of modifiable risk factors like smoking, obesity, physical inactivity, and poor nutrition.1,2 Other potentially modifiable causes include poverty, social inequality, poor education, and environmental factors.1,2 These wider influences are, again, rarely considered in clinical trials. Of relevance to Dr Leighton’s longstanding interest in the Maritimes is that this region has generally poorer health, lower incomes, higher unemployment, and a smaller proportionate share of the national wealth than the rest of Canada.
The region also has higher rates of smoking, obesity, and physical inactivity, which are both risk factors for and symptoms of socioeconomic inequity.3 As a result, Nova Scotia has the lowest disability-free life expectancy in the country, 3 years less than the Canadian average.4 Physical and mental illnesses are closely related, and administrative data can complement community surveys such as Stirling County, or the Canadian Community Health Survey, in exploring these issues further. While lacking the rich detail of community surveys, administrative data have advantages in terms of coverage and cost. They also can be continuously updated. For example, the Public Health Agency of Canada has developed case definitions for the surveillance of mental health disorders using administrative data across Canada.5 It is also possible to link data extracts across databases—for example, hospital morbidity and physician billings—using encrypted or anonymized identifiers. This offers the opportunity to track progress through the entire health system. Of particular relevance to chronic diseases, whose causes are often multifactorial, it is increasingly possible to link health data to income level, deprivation indices, and geospatial information. Studies of administrative data have shown that people with psychiatric disorders are twice as likely to be admitted for physical conditions which, with appropriate primary care, should not require inpatient treatment, including complications of diabetes, cardiovascular disorders, and acute
1
2
Departments of Psychiatry, Community Health and Epidemiology, Dalhousie University, Nova Scotia, Canada School of Medicine, University of Queensland, Queensland, Australia
Corresponding Author: Steve Kisely, MD, PhD, School of Medicine, the University of Queensland, Level 4, Building 1, Princess Alexandra Hospital, 199 Ipswich Rd, Woolloongabba, QLD 4102, Australia. Email: [email protected]
La Revue Canadienne de Psychiatrie 62(3)
183
Figure 1. Ratios of cancer incidence and mortality rates in male psychiatric patients compared with the general population in Nova Scotia.
exacerbation of respiratory conditions.6 Canadian data also show that people with schizophrenia have more physician contacts for diabetes and cardiovascular disease.7 But it is the increase in preventable and premature mortality in people with mental illness, particularly severe mental illness, that is of most concern. This was noted as far back as the early 1930s when mortality rates within the major asylums of New York were compared with mortality rates in the general population.8 The mortality rates in these institutions were increased for both males and females across the lifespan, with the major causes being conditions such as cardiovascular disease, not suicide. A similar pattern remains to this day. Patients with severe mental illnesses such as schizophrenia, bipolar, and delusional disorders die 15 to 20 years earlier than the general population.9 Excess mortality from chronic physical diseases is up to 10 times that of deliberate self-harm yet receives far less attention than suicide.9 Canadian studies also highlight that this not a problem restricted to patients of mental health services.10,11 In contrast to many studies elsewhere that are restricted to patients of specialised psychiatric services, studies from Canada often include people who are attending primary care for their mental health problems. These samples are therefore more representative of the general population given that the vast majority of psychiatric morbidity is treated by family doctors. Although the disparity in life expectancy is less than in the United States or Australia,10 possibly related to better universal health care and a more representative sample, 72% of excess deaths occurred in people who had only ever attended general practice for their psychiatric care.11 Furthermore, evidence indicates that the gap in mortality has been widening over the last few decades despite a shift to community-based treatment and the integration of psychiatric and general medical services.12,13 Some of these changes could have been expected to reduce the level of exposure to some mortality risks, such as infectious diseases, and improve access to physical health care.14 At the same time,
other risks could have increased with more freedom to make unhealthy lifestyle choices, or experience adverse social circumstances such as homelessness.14 This widening gap is especially true of cardiovascular disease, as shown by an increase in mortality in psychiatric patients at a time when deaths from ischaemic heart disease in the general population have been falling. People with mental illness have therefore not benefitted from preventative and public health measures, such a smoking cessation or the promotion of exercise, that have helped the general population. However, lifestyle factors such as alcohol, tobacco, and substance use, as well as diet and lack of exercise, are unlikely to be the sole answer. For instance, the incidence of many cancers, including melanoma, is no higher than in the general population while mortality is greater (Figure 1).15 If lifestyle were the only cause, incidence should more closely match mortality. Furthermore, concurrent lifestyle would not explain the findings for a cancer such as melanoma where childhood sun exposure is the main risk factor. Of course, it is possible that a lower than expected cancer incidence could be due to people with mental illness dying prematurely from lifestyle-induced cardiovascular disease. However, all-cause mortality remains high even after adjustment for behavioural risk factors such as smoking, physical activity, and body mass index.16 If people are no more likely to develop a disorder but are more likely to die of it, this suggests possible issues around treatment.17-19 These include delays in detection or initial presentation, leading to more advanced illness at diagnosis, or decreased access to services.17 Canadian and Australian studies have shown that psychiatric patients are more likely to die of cardiovascular disorders and cancer but less likely to receive the appropriate treatment such as coronary artery bypass grafts, surgery, or chemo/radiotherapy.17-22 Canadians with psychoses are also less likely to be discharged with guideline-consistent medication following a myocardial infarction.20
184
Administrative data have been used to study the long-term effects of poor oral health in people with mental illness. This is a particularly neglected area even though psychiatric patients have several relevant risk factors, including diet, lifestyle, and psychotropic-induced dry mouth. As a result, people with severe mental illnesses are 3 times as likely to have lost all their teeth compared with the general popluation.23 Poor oral health is more than just a question of aesthetics or pain; it can also predispose to chronic physical diseases such as myocardial infarction, stroke, and diabetes. For instance, studies of hospital administrative data have shown that dental disease accounts for 13% of avoidable admissions for preventable medical conditions in people with mental illness.6 Ways of addressing these problems include a greater awareness among clinicians of physical symptoms, registration with a family physician, and shared mental health care. However, many of these interventions are complex, and we also need to consider the influence of socioeconomic or environmental factors in determining response. Clinical trials are less suitable for this type of evaluation as they usually exclude patients with comorbidities, including alcohol or substance use. Epidemiological studies are therefore vital to understanding these wider environmental factors so as to enable population and health service–level interventions to address these multiple factors. In addition, by linking prescription data from provincial health plans, or pharmacies, to health service use and mortality, it may be possible to study the effects of medication in the ‘‘real world’’ throughout a drug’s life cycle.24 By contrast, Phase 3 clinical trials provide little information on long-term use, drug interactions, or high-risk, complex patients who may be more susceptible to adverse events, while Phase 4 trials are expensive to conduct. Continuous surveillance can, therefore, establish safety and effectiveness as well as detect rare, but serious, adverse events. But the uses of health administrative data extend beyond surveillance.25 Combined with advanced analytics, these data can also be used in the modelling and evaluation of interventions within complex health systems. For instance, these approaches can reveal causal relationships that would help in prediction, signal detection, and the assessment of real-world effectiveness across the diversity of patients in usual medical practice. Of course there are limitations. Administrative data are designed for billing, rather than evaluation, and both data quality and coverage may vary by province. Provincial drug plans are a notable example. Where information on data quality does exist, it primarily involves hospital admissions or separations rather than physician billings, where most treatment occurs. In addition, administrative databases capture data only on patients treated by medically trained practitioners, not other health professionals such as psychologists or occupational therapists. Data on all but the most severe cases of alcohol or substance use are rarely captured. The information is mostly restricted to demographic characteristics, health service contacts, and diagnosis. Important outcomes such as recovery, wellness, quality of life, or social
The Canadian Journal of Psychiatry 62(3)
determinants of health are rarely considered, nor are the views of patients. Finally, it can take researchers some time to obtain permission for access to the data. Nevertheless, administrative health data can both quantify the problem and track the effects of interventions at programme and population levels on a continuous basis at relatively little cost. These data cover patients such as those with multiple chronic comorbidities who are not typical of people in clinical trials. In the future, the utility of these data would be enhanced by linkage to medical and dental insurance plans, or workers’ compensation schemes, as well as data outside of health such as education. Big data such as these will play an ever-increasing role in epidemiology, and Canada is well placed to benefit from these developments. Author’s Note This article is a summary of a lecture delivered by Dr. Kisely, the 2015 recipient of the Alex Leighton Joint Canadian Psychiatric Association–Canadian Academy of Psychiatric Epidemiology (CPA-CAPE) Award in Psychiatric Epidemiology.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Australian Institute of Health and Welfare. Australia’s health 2014. Australia’s health series no. 14. Cat. no. AUS 178. Canberra: AIHW, 2014. 2. Anand P, Kunnumakara AB, Sundaram C, Harikumar KB, Tharakan ST, Lai OS, Sung B, Aggarwal BB. Cancer is a preventable disease that requires major lifestyle changes. Pharm Res. 2008;25(9):2097-2116. 3. Hayward K, Colman R. The tides of change: addressing inequity and chronic disease in Atlantic Canada: a discussion paper. Halifax (NS): Health Canada, Population and Public Health Branch, Atlantic Region; 2003. 4. Mirolla M. The cost of chronic disease in Canada. January 2004. Available from: http://www.gpiatlantic.org/pdf/health/ chroniccanada.pdf. Accessed April 14, 2016. 5. Kisely S, Lin E, Lesage A, Gilbert C, Smith M, Campbell LA, Vasiliadis HM. Use of administrative data for the surveillance of mental disorders in five provinces. Can J Psychiatry. 2009; 54(8):571-575. 6. Kisely S, Ehrlich C, Kendall E, Lawrence D. Using avoidable admissions to measure quality of care for cardiometabolic and other physical comorbidities of psychiatric disorders: a population-based record-linkage analysis. Can J Psychiatry. 2015;60(11):497-506. 7. Bresee LC, Majumdar SR, Patten SB, Johnson JA. Prevalence of cardiovascular risk factors and disease in people with
La Revue Canadienne de Psychiatrie 62(3)
8. 9.
10.
11.
12.
13.
14.
15.
16.
schizophrenia: a population-based study. Schizophr Res. 2010; 117(1):75-82. Malzberg B. Life tables for patients with mental disease. Psychiatr Q. 1932;6(2):226-241. Lawrence D, Hancock K, Kisely S. The gap in life expectancy from preventable physical illness in psychiatric patients in Western Australia: retrospective analysis of population based registers. BMJ. 2013;346:f2539. Lesage A, Rochette L, Emond V, Pelletier E, St-Laurent D, Diallo F, Kisely S. A surveillance system to monitor the excess mortality of the mentally ill in the Canadian context. Can J Psychiatry. 2015;60(12):571-579. Kisely S, Smith M, Lawrence D, Maaten S. Mortality in individuals who have had psychiatric treatment in Nova Scotia. Br J Psychiatry. 2005;187:552-558. Saha S, Chant D, McGrath J. A systematic review of mortality in schizophrenia: is the differential mortality gap worsening over time? Arch Gen Psychiatry. 2007;64(10):1123-1131. Lawrence DM, Holman CD, Jablensky AV, Hobbs MS. Death rate from ischaemic heart disease in Western Australian psychiatric patients 1980-1998. Br J Psychiatry. 2003;182:31-36. Kasprow WJ, Rosenheck R. Mortality among homeless and non-homeless mentally ill veterans. J Nerv Ment Dis. 2000; 188:141-147. Kisely S, Sadek J, MacKenzie A, Lawrence D, Campbell LA. Excess cancer mortality in psychiatric patients. Can J Psychiatry. 2008;53(11):753-761. Hamer M, Stamatakis E, Steptoe A. Psychiatric hospital admissions, behavioral risk factors, and all-cause mortality: the Scottish health survey. Arch Intern Med. 2008;168:2474-2479.
185
17. Kisely S, Crowe E, Lawrence D. Assessing why people with mental illness are no more likely than the general population to develop cancer but more likely to die of it. JAMA Psychiatry. 2013;70(2):209-217. 18. Kisely S, Cox M, Smith M, Lawrence D, Maaten S. Does inequitable access to cardiological or neurological procedures contribute to preventable mortality in people with mental illness? Can Med Assoc J. 2007;176:779-784. 19. Druss BG, Bradford DW, Rosenheck RA, Radford MJ, Krumholz HM. Mental disorders and use of cardiovascular procedures after myocardial infarction. JAMA. 2000;283(4):506-511. 20. Kisely S, Campbell LA, Wang Y. Treatment of ischaemic heart disease and stroke in individuals with psychosis under universal healthcare. Br J Psychiatry. 2009;195(6):545-550. 21. Bresee LC, Majumdar SR, Patten SB, Johnson JA. Utilization of general and specialized cardiac care by people with schizophrenia. Psychiatr Serv. 2012;63(3):237-242. 22. Kurdyak P, Vigod S, Calzavara A, Wodchis WP. High mortality and low access to care following incident acute myocardial infarction in individuals with schizophrenia. Schizophr Res. 2012;142(1-3):52-57. 23. Kisely S. No mental health without oral health. Can J Psychiatry. 2016;61(5):277-282. 24. Kisely S, Cox M, Campbell LA, Cooke C, Gardner D. An epidemiological study of psychotropic medication and obesity-related chronic illnesses in older psychiatric patients. Can J Psychiatry. 2009;54(4):269-274. 25. Krumholz HM. Big data and new knowledge in medicine: the thinking, training, and tools needed for a learning health system. Health Aff (Millwood). 2014;33(7):1163-1170.
Association des psychiatres du Canada
Perspective
On Adjusting for Life’s Confounding: Harnessing Big Data to Answer Big Problems
The Canadian Journal of Psychiatry / La Revue Canadienne de Psychiatrie 2017, Vol. 62(3) 182-185 ª The Author(s) 2016 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0706743716660710 TheCJP.ca | LaRCP.ca
Steve Kisely, MD, PhD1,2
Keywords epidemiology, surveillance, mental health services, health services research, comorbidity
A wide range of data sources are available to study the epidemiology of mental illness. Community surveys such as Alex Leighton’s Stirling County Study are the gold standard for such data sources, particularly because community surveys cover everyone, not just people seeking treatment. However, as surveys require a lot of resources, other methods are also used, each of which has strengths and weaknesses. For example, medical records contain detailed information, but this can be difficult to extract and the quality may vary. Another source entails administrative data, typically hospital separations, physician billings, ambulatory care visits, and drug databases. While such data require careful analysis with the use of multivariate or propensity score techniques to adjust for potential confounding variables, these data can be invaluable in the study of diseases with multifactorial aetiologies. This paper addresses how these large administrative databases, or big data, can help address mental health problems in situations where conventional clinical studies are of little help. Patients with chronic disease are an example, given they are often excluded from clinical trials on account of their multiple medical and psychiatric comorbidities. This is an important omission, as these diseases are major sources of disability and burden worldwide. While the causes of chronic illnesses are not fully known, a range of apparently interrelated personal, social, economic, and environmental factors are associated with their development. As much as 40% of chronic illness may be attributable to a small number of modifiable risk factors like smoking, obesity, physical inactivity, and poor nutrition.1,2 Other potentially modifiable causes include poverty, social inequality, poor education, and environmental factors.1,2 These wider influences are, again, rarely considered in clinical trials. Of relevance to Dr Leighton’s longstanding interest in the Maritimes is that this region has generally poorer health, lower incomes, higher unemployment, and a smaller proportionate share of the national wealth than the rest of Canada.
The region also has higher rates of smoking, obesity, and physical inactivity, which are both risk factors for and symptoms of socioeconomic inequity.3 As a result, Nova Scotia has the lowest disability-free life expectancy in the country, 3 years less than the Canadian average.4 Physical and mental illnesses are closely related, and administrative data can complement community surveys such as Stirling County, or the Canadian Community Health Survey, in exploring these issues further. While lacking the rich detail of community surveys, administrative data have advantages in terms of coverage and cost. They also can be continuously updated. For example, the Public Health Agency of Canada has developed case definitions for the surveillance of mental health disorders using administrative data across Canada.5 It is also possible to link data extracts across databases—for example, hospital morbidity and physician billings—using encrypted or anonymized identifiers. This offers the opportunity to track progress through the entire health system. Of particular relevance to chronic diseases, whose causes are often multifactorial, it is increasingly possible to link health data to income level, deprivation indices, and geospatial information. Studies of administrative data have shown that people with psychiatric disorders are twice as likely to be admitted for physical conditions which, with appropriate primary care, should not require inpatient treatment, including complications of diabetes, cardiovascular disorders, and acute
1
2
Departments of Psychiatry, Community Health and Epidemiology, Dalhousie University, Nova Scotia, Canada School of Medicine, University of Queensland, Queensland, Australia
Corresponding Author: Steve Kisely, MD, PhD, School of Medicine, the University of Queensland, Level 4, Building 1, Princess Alexandra Hospital, 199 Ipswich Rd, Woolloongabba, QLD 4102, Australia. Email: [email protected]
La Revue Canadienne de Psychiatrie 62(3)
183
Figure 1. Ratios of cancer incidence and mortality rates in male psychiatric patients compared with the general population in Nova Scotia.
exacerbation of respiratory conditions.6 Canadian data also show that people with schizophrenia have more physician contacts for diabetes and cardiovascular disease.7 But it is the increase in preventable and premature mortality in people with mental illness, particularly severe mental illness, that is of most concern. This was noted as far back as the early 1930s when mortality rates within the major asylums of New York were compared with mortality rates in the general population.8 The mortality rates in these institutions were increased for both males and females across the lifespan, with the major causes being conditions such as cardiovascular disease, not suicide. A similar pattern remains to this day. Patients with severe mental illnesses such as schizophrenia, bipolar, and delusional disorders die 15 to 20 years earlier than the general population.9 Excess mortality from chronic physical diseases is up to 10 times that of deliberate self-harm yet receives far less attention than suicide.9 Canadian studies also highlight that this not a problem restricted to patients of mental health services.10,11 In contrast to many studies elsewhere that are restricted to patients of specialised psychiatric services, studies from Canada often include people who are attending primary care for their mental health problems. These samples are therefore more representative of the general population given that the vast majority of psychiatric morbidity is treated by family doctors. Although the disparity in life expectancy is less than in the United States or Australia,10 possibly related to better universal health care and a more representative sample, 72% of excess deaths occurred in people who had only ever attended general practice for their psychiatric care.11 Furthermore, evidence indicates that the gap in mortality has been widening over the last few decades despite a shift to community-based treatment and the integration of psychiatric and general medical services.12,13 Some of these changes could have been expected to reduce the level of exposure to some mortality risks, such as infectious diseases, and improve access to physical health care.14 At the same time,
other risks could have increased with more freedom to make unhealthy lifestyle choices, or experience adverse social circumstances such as homelessness.14 This widening gap is especially true of cardiovascular disease, as shown by an increase in mortality in psychiatric patients at a time when deaths from ischaemic heart disease in the general population have been falling. People with mental illness have therefore not benefitted from preventative and public health measures, such a smoking cessation or the promotion of exercise, that have helped the general population. However, lifestyle factors such as alcohol, tobacco, and substance use, as well as diet and lack of exercise, are unlikely to be the sole answer. For instance, the incidence of many cancers, including melanoma, is no higher than in the general population while mortality is greater (Figure 1).15 If lifestyle were the only cause, incidence should more closely match mortality. Furthermore, concurrent lifestyle would not explain the findings for a cancer such as melanoma where childhood sun exposure is the main risk factor. Of course, it is possible that a lower than expected cancer incidence could be due to people with mental illness dying prematurely from lifestyle-induced cardiovascular disease. However, all-cause mortality remains high even after adjustment for behavioural risk factors such as smoking, physical activity, and body mass index.16 If people are no more likely to develop a disorder but are more likely to die of it, this suggests possible issues around treatment.17-19 These include delays in detection or initial presentation, leading to more advanced illness at diagnosis, or decreased access to services.17 Canadian and Australian studies have shown that psychiatric patients are more likely to die of cardiovascular disorders and cancer but less likely to receive the appropriate treatment such as coronary artery bypass grafts, surgery, or chemo/radiotherapy.17-22 Canadians with psychoses are also less likely to be discharged with guideline-consistent medication following a myocardial infarction.20
184
Administrative data have been used to study the long-term effects of poor oral health in people with mental illness. This is a particularly neglected area even though psychiatric patients have several relevant risk factors, including diet, lifestyle, and psychotropic-induced dry mouth. As a result, people with severe mental illnesses are 3 times as likely to have lost all their teeth compared with the general popluation.23 Poor oral health is more than just a question of aesthetics or pain; it can also predispose to chronic physical diseases such as myocardial infarction, stroke, and diabetes. For instance, studies of hospital administrative data have shown that dental disease accounts for 13% of avoidable admissions for preventable medical conditions in people with mental illness.6 Ways of addressing these problems include a greater awareness among clinicians of physical symptoms, registration with a family physician, and shared mental health care. However, many of these interventions are complex, and we also need to consider the influence of socioeconomic or environmental factors in determining response. Clinical trials are less suitable for this type of evaluation as they usually exclude patients with comorbidities, including alcohol or substance use. Epidemiological studies are therefore vital to understanding these wider environmental factors so as to enable population and health service–level interventions to address these multiple factors. In addition, by linking prescription data from provincial health plans, or pharmacies, to health service use and mortality, it may be possible to study the effects of medication in the ‘‘real world’’ throughout a drug’s life cycle.24 By contrast, Phase 3 clinical trials provide little information on long-term use, drug interactions, or high-risk, complex patients who may be more susceptible to adverse events, while Phase 4 trials are expensive to conduct. Continuous surveillance can, therefore, establish safety and effectiveness as well as detect rare, but serious, adverse events. But the uses of health administrative data extend beyond surveillance.25 Combined with advanced analytics, these data can also be used in the modelling and evaluation of interventions within complex health systems. For instance, these approaches can reveal causal relationships that would help in prediction, signal detection, and the assessment of real-world effectiveness across the diversity of patients in usual medical practice. Of course there are limitations. Administrative data are designed for billing, rather than evaluation, and both data quality and coverage may vary by province. Provincial drug plans are a notable example. Where information on data quality does exist, it primarily involves hospital admissions or separations rather than physician billings, where most treatment occurs. In addition, administrative databases capture data only on patients treated by medically trained practitioners, not other health professionals such as psychologists or occupational therapists. Data on all but the most severe cases of alcohol or substance use are rarely captured. The information is mostly restricted to demographic characteristics, health service contacts, and diagnosis. Important outcomes such as recovery, wellness, quality of life, or social
The Canadian Journal of Psychiatry 62(3)
determinants of health are rarely considered, nor are the views of patients. Finally, it can take researchers some time to obtain permission for access to the data. Nevertheless, administrative health data can both quantify the problem and track the effects of interventions at programme and population levels on a continuous basis at relatively little cost. These data cover patients such as those with multiple chronic comorbidities who are not typical of people in clinical trials. In the future, the utility of these data would be enhanced by linkage to medical and dental insurance plans, or workers’ compensation schemes, as well as data outside of health such as education. Big data such as these will play an ever-increasing role in epidemiology, and Canada is well placed to benefit from these developments. Author’s Note This article is a summary of a lecture delivered by Dr. Kisely, the 2015 recipient of the Alex Leighton Joint Canadian Psychiatric Association–Canadian Academy of Psychiatric Epidemiology (CPA-CAPE) Award in Psychiatric Epidemiology.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Australian Institute of Health and Welfare. Australia’s health 2014. Australia’s health series no. 14. Cat. no. AUS 178. Canberra: AIHW, 2014. 2. Anand P, Kunnumakara AB, Sundaram C, Harikumar KB, Tharakan ST, Lai OS, Sung B, Aggarwal BB. Cancer is a preventable disease that requires major lifestyle changes. Pharm Res. 2008;25(9):2097-2116. 3. Hayward K, Colman R. The tides of change: addressing inequity and chronic disease in Atlantic Canada: a discussion paper. Halifax (NS): Health Canada, Population and Public Health Branch, Atlantic Region; 2003. 4. Mirolla M. The cost of chronic disease in Canada. January 2004. Available from: http://www.gpiatlantic.org/pdf/health/ chroniccanada.pdf. Accessed April 14, 2016. 5. Kisely S, Lin E, Lesage A, Gilbert C, Smith M, Campbell LA, Vasiliadis HM. Use of administrative data for the surveillance of mental disorders in five provinces. Can J Psychiatry. 2009; 54(8):571-575. 6. Kisely S, Ehrlich C, Kendall E, Lawrence D. Using avoidable admissions to measure quality of care for cardiometabolic and other physical comorbidities of psychiatric disorders: a population-based record-linkage analysis. Can J Psychiatry. 2015;60(11):497-506. 7. Bresee LC, Majumdar SR, Patten SB, Johnson JA. Prevalence of cardiovascular risk factors and disease in people with
La Revue Canadienne de Psychiatrie 62(3)
8. 9.
10.
11.
12.
13.
14.
15.
16.
schizophrenia: a population-based study. Schizophr Res. 2010; 117(1):75-82. Malzberg B. Life tables for patients with mental disease. Psychiatr Q. 1932;6(2):226-241. Lawrence D, Hancock K, Kisely S. The gap in life expectancy from preventable physical illness in psychiatric patients in Western Australia: retrospective analysis of population based registers. BMJ. 2013;346:f2539. Lesage A, Rochette L, Emond V, Pelletier E, St-Laurent D, Diallo F, Kisely S. A surveillance system to monitor the excess mortality of the mentally ill in the Canadian context. Can J Psychiatry. 2015;60(12):571-579. Kisely S, Smith M, Lawrence D, Maaten S. Mortality in individuals who have had psychiatric treatment in Nova Scotia. Br J Psychiatry. 2005;187:552-558. Saha S, Chant D, McGrath J. A systematic review of mortality in schizophrenia: is the differential mortality gap worsening over time? Arch Gen Psychiatry. 2007;64(10):1123-1131. Lawrence DM, Holman CD, Jablensky AV, Hobbs MS. Death rate from ischaemic heart disease in Western Australian psychiatric patients 1980-1998. Br J Psychiatry. 2003;182:31-36. Kasprow WJ, Rosenheck R. Mortality among homeless and non-homeless mentally ill veterans. J Nerv Ment Dis. 2000; 188:141-147. Kisely S, Sadek J, MacKenzie A, Lawrence D, Campbell LA. Excess cancer mortality in psychiatric patients. Can J Psychiatry. 2008;53(11):753-761. Hamer M, Stamatakis E, Steptoe A. Psychiatric hospital admissions, behavioral risk factors, and all-cause mortality: the Scottish health survey. Arch Intern Med. 2008;168:2474-2479.
185
17. Kisely S, Crowe E, Lawrence D. Assessing why people with mental illness are no more likely than the general population to develop cancer but more likely to die of it. JAMA Psychiatry. 2013;70(2):209-217. 18. Kisely S, Cox M, Smith M, Lawrence D, Maaten S. Does inequitable access to cardiological or neurological procedures contribute to preventable mortality in people with mental illness? Can Med Assoc J. 2007;176:779-784. 19. Druss BG, Bradford DW, Rosenheck RA, Radford MJ, Krumholz HM. Mental disorders and use of cardiovascular procedures after myocardial infarction. JAMA. 2000;283(4):506-511. 20. Kisely S, Campbell LA, Wang Y. Treatment of ischaemic heart disease and stroke in individuals with psychosis under universal healthcare. Br J Psychiatry. 2009;195(6):545-550. 21. Bresee LC, Majumdar SR, Patten SB, Johnson JA. Utilization of general and specialized cardiac care by people with schizophrenia. Psychiatr Serv. 2012;63(3):237-242. 22. Kurdyak P, Vigod S, Calzavara A, Wodchis WP. High mortality and low access to care following incident acute myocardial infarction in individuals with schizophrenia. Schizophr Res. 2012;142(1-3):52-57. 23. Kisely S. No mental health without oral health. Can J Psychiatry. 2016;61(5):277-282. 24. Kisely S, Cox M, Campbell LA, Cooke C, Gardner D. An epidemiological study of psychotropic medication and obesity-related chronic illnesses in older psychiatric patients. Can J Psychiatry. 2009;54(4):269-274. 25. Krumholz HM. Big data and new knowledge in medicine: the thinking, training, and tools needed for a learning health system. Health Aff (Millwood). 2014;33(7):1163-1170.
