Diabetes in Ontario
Practice Atlas
1 Chapter
Patterns of Prevalence and Incidence of Diabetes Authors: Janet E. Hux and Mei Tang
1.1
Diabetes in Ontario
Patterns of Prevalence and Incidence of Diabetes
Background Diabetes mellitus (DM) is a common, chronic condition that imposes a heavy burden of morbidity (illness) and early mortality (death) on affected patients.1-3 DM and its complications drive a substantial portion of medical resource utilization. At the same time, research findings now provide unprecedented levels of evidence regarding the prevention of DM complications.4-9 In this context, accurate, population-based assessments of the prevalence of DM become important for policy-makers and for those mounting and evaluating strategies for managing this condition.
Key Messages • Diabetes mellitus (DM) is a large and growing health problem for Ontarians.
• Primary care providers can expect to deal with increased numbers of patients with DM, patients who are living longer and will have more advanced stages of disease.
• The high prevalence of DM in the elderly has important implications for health care resource utilization given the burden of DM and the projected growth of this segment of the population.
• Effective management of DM in older persons is critical, making it important to include individuals in this age group in clinical trials.
• Providers need to be aware of the ethnic, geographic and socioeconomic factors that increase the risk of DM. Strategies to address issues related to access, prevention, and treatment of individuals in these high-risk groups are needed.
Evidence from other jurisdictions suggests that the prevalence of DM is rising.10-13 Prevalence reflects the total number of persons in a population with DM at a given point in time—both those newly diagnosed and those already living with the condition. Prevalence may increase because there are growing numbers of new cases entering the population each year, because those diagnosed with the condition are living longer, or both. An increase in the number of incident cases (persons newly diagnosed with DM) might be expected given the rising rates of obesity10 and changing demographics. Improvement in survival might be anticipated because of the increasing availability of effective interventions for the prevention and control of DM complications. There is also the possibility that earlier detection of DM, or changes in the threshold for diagnosis, might create the impression that the incidence of DM has increased. There is a lack of consensus about the most effective way of determining the prevalence of DM in a population. Previous work has based prevalence estimates on surveys,14-16 registries17 and cohort studies in highly selected populations.18 Health interview programs such as the National Population Health Survey (NPHS) have facilitated population-based estimates. However, there is evidence that in health interview surveys (i.e. where no blood samples are obtained) participants under-report DM relative to medical record reviews.19, 20 Surveys suffer from biases due to low response rates, providing insufficient data to define prevalence at the level of small geographic areas and are inefficient for ongoing surveillance. Research by Blanchard and colleagues in Manitoba21 showed that health care administrative data can be used to identify individuals diagnosed with DM in the province and to estimate rates over time. Their methodology has been adopted by the National Diabetes Surveillance System (NDSS). The NDSS is a Health Canada initiative involving provinces and territories in DM surveillance, using administrative data to conduct analyses based on common guidelines and software. In this way, the data can be meaningfully aggregated to provide a national profile of DM. Prior to the implementation of the NDSS in Ontario, researchers at the Institute for Clinical Evaluative Sciences (ICES) had developed a provincial database, the Ontario Diabetes Database (ODD), using algorithms similar to those developed for the NDSS. The development and validation of the ODD is described in the Technical Appendix TA1.A.
1.2
Practice Atlas
Exhibit 1.1 Overall and Age-/Sex-specific DM Prevalence Rates per 100 Ontarians, 1995–1999 The prevalence of DM rises with age and is generally higher in men than in women. Prevalence rates increased steadily over the years that were studied. Fiscal Year
Overall Rate
20–34
35–49
Women by Age Group 50–64
65–74
75+
Men by Age Group
1995
4.72
0.79
2.20
6.84
11.58
12.58
1996
5.09
0.84
2.41
7.24
12.31
1997
5.45
0.90
2.57
7.64
13.10
1998
5.82
0.96
2.77
8.04
1999
6.19
1.02
2.97
8.40
20–34
35–49
50–64
65–74
75+
0.65
2.78
9.15
14.75
15.75
13.51
0.69
2.98
9.79
15.82
16.93
14.36
0.72
3.13
10.43
16.80
17.91
13.89
15.17
0.74
3.28
10.99
17.74
18.98
14.62
15.97
0.77
3.44
11.50
18.69
20.09
Source: Ontario Diabetes Database (ODD)
This chapter provides an indication of the magnitude of the burden of DM in Ontario. It describes how the patterns of DM are changing. It further explores its distribution across geographic regions, as well as by age, sex, and socioeconomic groupings.
Data Sources The major source of data for this chapter is the Ontario Diabetes Database (ODD). This database was prepared at ICES using hospital discharge abstracts from the Canadian Institute for Health Information (CIHI), physician service claims from the Ontario Health Insurance Plan (OHIP) database and information regarding the demographics of persons eligible for health care coverage in Ontario from the Registered Persons Database (RPDB). Records from these three sources for all persons in Ontario were linked using an anonymous numeric identifier. Persons were defined as having DM (excluding cases of gestational diabetes) according to criteria described in the Technical Appendix TA1.A. Claims to the Ontario Drug Benefit (ODB) Program were used for validation. Census data from Statistics Canada were used to establish denominators for calculation of DM rates and to attribute socioeconomic characteristics to the forward sortation area (or local neighbourhood).
How the analysis was done Prevalence is the proportion of the population affected by a condition at a given point in time. Prevalence of DM was calculated on an annual basis from fiscal 1995 (April 1, 1994 to March 31, 1995) through fiscal 2000 using all persons in the ODD for each year as the numerator and census counts for the population as the denominator (or estimated population measures for those years where there was no census). To adjust for differences in population distribution over time, rates were ageand sex-adjusted to the 1996 Ontario population using direct standardization. Incidence rates were calculated in a similar fashion using only the incident cases for a given year as the numerator. The ODD data are available for fiscal years 1992
1.3
1
Diabetes in Ontario
Patterns of Prevalence and Incidence of Diabetes
Exhibit 1.2 Overall and Age-/Sex-specific DM Incidence Rates per 100 Ontarians, 1995–1999 Incidence rates (persons newly diagnosed with DM) increase with age and are generally higher in men than women. In contrast to prevalence rates, the incidence rates appear to be stable over the years studied. Fiscal Year
Overall Rate
20–34
35–49
Women by Age Group 50–64
65–74
75+
20–34
35–49
Men by Age Group 50–64
65–74
75+
1995
0.68
0.15
0.40
1.00
1.38
1.40
0.12
0.54
1.38
1.82
1.77
1996
0.62
0.14
0.38
0.92
1.24
1.25
0.12
0.49
1.27
1.65
1.52
1997
0.61
0.15
0.37
0.91
1.26
1.24
0.11
0.47
1.26
1.56
1.49
1998
0.66
0.17
0.41
0.99
1.32
1.28
0.12
0.51
1.30
1.66
1.56
1999
0.66
0.18
0.41
0.95
1.28
1.25
0.13
0.51
1.28
1.65
1.56
Source: Ontario Diabetes Database (ODD)
through 2000. In order to identify an incident case (newly diagnosed), a minimum DM-free observation period of three years was set as a requirement. For example, a person meeting the criteria for entering the database in 1995 must have had no OHIP or CIHI records bearing a diagnosis of DM during the previous three years to be labeled as an “incident” case. As a result, the incidence of DM prior to 1995 could not be estimated because a three-year pre-diagnosis observation period was not available. Socioeconomic status (SES) is known to be an important factor in the epidemiology of DM. However, data on SES are not reported at an individual person level in the available administrative data files. Therefore, in order to estimate the SES of persons with DM, the neighbourhood level median household income from census data was attributed to all persons living in that neighbourhood. Neighbourhood of residence was determined from the postal code in the RPDB and matched to census data at the level of the forward sortation area. The median population of these units in the 1996 census was 19,000 persons. Rates and numbers of cases of DM were also calculated at the county level.
Interpretative Cautions Administrative data provide imperfect estimates of the rates of DM. At best these data can only be used to measure rates of diagnosed DM and are unable to provide estimates of undiagnosed DM. Studies in other jurisdictions suggest that up to 30% of DM may be undiagnosed.22 In addition, persons with diagnosed DM may not be detected by the algorithm used here if they receive their care in a setting where services are not billed on a fee-for-service basis. This pattern of service represents only a small proportion of primary care ( 3 office visits coded with diagnosis of DM
55 (64.7)
Persons with at least 1 hospitalization with a diagnosis of DM
11 (12.9)
Persons having seen more than 3 different physicians in last 5 years*
63 (74.1)
Persons with one or more of the above 4 criteria
78 (91.8)
* Where a patient routinely sees multiple physicians, it is less likely that a given physician (i.e. the one whose charts were abstracted) would have the patient’s full medical history including the diagnosis of DM. Sources: Canadian Institute for Health Information (CIHI), Ontario Drug Benefit Plan (ODB), Ontario Health Insurance Plan (OHIP)
References: 1.
Blanchard JF, Ludwig S, Wajda A et al. Incidence and prevalence of diabetes in Manitoba, 1986–1991. Diabetes Care1996; 19:807–811.
2.
Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: Determination of Prevalence and Incidence Using a Validated Administrative Data Algorithm. Diabetes Care 2002; 25(3):512–516.
1.18
