Advertisement

Past and Future Burden of Inflammatory Bowel Diseases Based on Modeling of Population-Based Data

Open AccessPublished:January 10, 2019DOI:https://doi.org/10.1053/j.gastro.2019.01.002

      Background & Aims

      Inflammatory bowel diseases (IBDs) exist worldwide, with high prevalence in North America. IBD is complex and costly, and its increasing prevalence places a greater stress on health care systems. We aimed to determine the past current, and future prevalences of IBD in Canada.

      Methods

      We performed a retrospective cohort study using population-based health administrative data from Alberta (2002–2015), British Columbia (1997–2014), Manitoba (1990–2013), Nova Scotia (1996–2009), Ontario (1999–2014), Quebec (2001–2008), and Saskatchewan (1998–2016). Autoregressive integrated moving average regression was applied, and prevalence, with 95% prediction intervals (PIs), was forecasted to 2030. Average annual percentage change, with 95% confidence intervals, was assessed with log binomial regression.

      Results

      In 2018, the prevalence of IBD in Canada was estimated at 725 per 100,000 (95% PI 716–735) and annual average percent change was estimated at 2.86% (95% confidence interval 2.80%–2.92%). The prevalence in 2030 was forecasted to be 981 per 100,000 (95% PI 963–999): 159 per 100,000 (95% PI 133–185) in children, 1118 per 100,000 (95% PI 1069–1168) in adults, and 1370 per 100,000 (95% PI 1312–1429) in the elderly. In 2018, 267,983 Canadians (95% PI 264,579–271,387) were estimated to be living with IBD, which was forecasted to increase to 402,853 (95% PI 395,466–410,240) by 2030.

      Conclusion

      Forecasting prevalence will allow health policy makers to develop policy that is necessary to address the challenges faced by health systems in providing high-quality and cost-effective care.

      Keywords

      Abbreviations used in this paper:

      AB (Alberta), AR (autoregressive), ARIMA (autoregressive integrated moving average), BC (British Columbia), CD (Crohn disease), CI (confidence interval), IBD (inflammatory bowel disease), MA (moving average), MB (Manitoba), NS (Nova Scotia), ON (Ontario), PI (prediction interval), QC (Quebec), SK (Saskatchewan), UC (ulcerative colitis)

       Background and Context

      The prevalence of Inflammatory Bowel Disease (IBD) in Canada is among the highest in the world and is steadily increasing. However, it is unclear how much the prevalence will increase in the future.

       New Findings

      In forecasting models, the prevalence of IBD, Crohn’s disease, and ulcerative colitis are significantly increasing – with some of the highest rates of increase seen within the pediatric and the elderly populations.

       Limitations

      This comprehensive analysis encompassing of 95% of the Canadian population lacks the ability to adjust for possible risk factors (eg, smoking) that affect the risk of developing IBD.

       Impact

      We are at a time where defining the future prevalence of disease – to 2030 – is of paramount importance as it will allow us to prepare for the impending burden.
      The prevalence of inflammatory bowel disease (IBD), comprised of Crohn disease (CD) and ulcerative colitis (UC), has been increasing globally, with one of the highest prevalence rates found in Canada.
      • Molodecky N.A.
      • Soon I.S.
      • Rabi D.M.
      • et al.
      Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review.
      • Ng S.C.
      • Shi H.Y.
      • Hamidi N.
      • et al.
      Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies.
      • Kaplan G.G.
      • Ng S.C.
      Globalisation of inflammatory bowel disease: perspectives from the evolution of inflammatory bowel disease in the UK and China.
      • Benchimol E.I.
      • Fortinsky K.J.
      • Gozdyra P.
      • et al.
      Epidemiology of pediatric inflammatory bowel disease: a systematic review of international trends.
      The incidence of IBD in the latter half of the 20th century increased significantly in the Western world, which has caused the prevalence to exceed 0.5% in North America.
      • Ng S.C.
      • Shi H.Y.
      • Hamidi N.
      • et al.
      Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies.
      The steadily rising prevalence of IBD can be attributed to the disease being diagnosed predominantly in young individuals, being chronic and incurable, and having low mortality.
      • Rocchi A.
      • Benchimol E.I.
      • Bernstein C.N.
      • et al.
      Inflammatory bowel disease: a Canadian burden of illness review.
      With incidence outpacing death, patients with newly diagnosed IBD are continually added to the pool of prevalent patients, leading to the compounding prevalence of IBD over time.
      • Kaplan G.G.
      The global burden of IBD: from 2015 to 2025.
      IBD is a complex and costly disease owing to an unpredictable relapsing and remitting course, complications, hospitalizations, surgeries, and use of expensive therapies. Thus, the steady increase in the prevalence of IBD will lead to a substantial increase in the burden borne by health care systems and society.
      • Kaplan G.G.
      The global burden of IBD: from 2015 to 2025.
      The present study aimed to forecast the future prevalence of IBD in Canada. The Canadian population is ideal for this study because Canada has a very high prevalence of IBD and places a large burden on the single-payer health care system, making it even more important to forecast future changes. Estimating the total number of people diagnosed with IBD will allow health care providers to proactively implement clinical practices and policy interventions to address the impact of the increasing prevalence of IBD.

      Methods

       Study Population and Data Sources

      The Canadian Gastrointestinal Epidemiology Consortium (CanGIEC) is a national collaboration of provincial IBD surveillance cohorts derived from administrative health care databases (Appendix 1). CanGIEC provided retrospective population-based provincial prevalence data of all individuals who qualify for health care, of any age, for Alberta (AB; 2002–2015), British Columbia (BC; 1997–2014), Manitoba (MB; 1990–2013), Nova Scotia (NS; 1996–2009), Ontario (ON; 1999–2014), Quebec (QC; 2001–2008), and Saskatchewan (SK; 1998–2016). These provinces account for approximately 95% of the Canadian population.

      Statistics Canada. Estimates of population, by age group and sex for July 1, Canada, provinces and territories annual. Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0510001&tabMode=dataTable&srchLan=-1&p1=-1&p2=9. Published 2017. Accessed June 5, 2017.

       Provincial Estimates

      The prevalence of IBD, CD, and UC were calculated for each province based on years of available data from the provincial administrative health care databases listed earlier using population values from Statistics Canada (eg, AB historical analysis was from 2002 to 2015 and forecasted from 2016 to 2030).

      Statistics Canada. Estimates of population, by age group and sex for July 1, Canada, provinces and territories annual. Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0510001&tabMode=dataTable&srchLan=-1&p1=-1&p2=9. Published 2017. Accessed June 5, 2017.

      Prevalence rates for each year were standardized for age and sex based on the Canadian population for the associated year.

      Statistics Canada. Estimates of population, by age group and sex for July 1, Canada, provinces and territories annual. Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0510001&tabMode=dataTable&srchLan=-1&p1=-1&p2=9. Published 2017. Accessed June 5, 2017.

       National Estimates

      Canadian population models were calculated by combining the prevalence for the 7 provinces together from 2002 to 2008—these were the only overlapping years when data were available from all provinces. Secondary analyses included sex stratification (male and female) and age stratification into pediatric (AB, BC, MB, ON, QC, and SK: 0–17 years; NS: 0–19 years), adult (AB, BC, MB, ON, QC, and SK: 18–59 years; NS: 20–59 years), and elderly (all provinces: ≥60 years) groups.

       Statistical Analysis

      The primary forecasting analysis was completed using autoregressive integrated moving average (ARIMA) models.
      • Bujang M.A.
      • Adnan T.H.
      • Hashim N.H.
      • et al.
      Forecasting the incidence and prevalence of patients with end-stage renal disease in Malaysia up to the year 2040.

      Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.
      An ARIMA model is a time series model in which the value being assessed is related back to historical values from periods before the current one; this is a method that has been used in econometric and prevalence forecasting.
      • Bujang M.A.
      • Adnan T.H.
      • Hashim N.H.
      • et al.
      Forecasting the incidence and prevalence of patients with end-stage renal disease in Malaysia up to the year 2040.

      Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.

      Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      This model was chosen because of its ability to analyze data at a specific period and relate those back to data contained in prior periods while accounting for innate dependence in yearly prevalence data. An ARIMA model is used for the analysis of equidistant, and discrete samples of data in a time series model, which is necessary with the analysis of prevalence data.

      Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.
      The ARIMA is composed of 3 components (1) autoregressive (AR), model (2) integrated, term and (3) moving average (MA) model.

      Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.

      Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      The AR model relates the value in a period to historical ones; the integrated term deals with the assumption of stationarity (that the probability distribution, mean and variance, remains constant over time); and the MA model relates the value back to historical residuals.

      Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.

      Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      First, the AR term, which integrates the prior values into the current one, can be evaluated using the partial autocorrelation function, delineating the possible lag periods that are to be included in the AR model. There can be more than 1 possibility, and the appropriate AR term is chosen in conjunction with the MA value. Second, the integrated term deals with the stationarity aspect of the data, which is evaluated using the Dickey-Fuller test.
      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.

      Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      If there is a lack of stationarity, then differencing is undertaken in an attempt to achieve stationarity.

      Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      In the present study, when differencing did not achieve stationarity, a log binomial model was used. Log binomial models analyze dichotomous outcomes, similar to logistic regression, but report outcomes as risk rather than odds. With the exception of the pediatrics and ON groups, all populations achieved stationarity, although the elderly group required differencing to achieve stationarity. Third, the MA term is evaluated with an autocorrelation function and, similar to the AR term, can have multiple possible values. Although the AR term and MA term have multiple possible values, and therefore numerous combinations of values, individual models of each possible value are created and evaluated for best fit. The best fitting models were chosen using a combination of the lowest Akaike information criterion, Bayesian information criterion, and root mean square error from each AR–MA combination. If a clear delineation between 2 models could not be made based on these criteria, then a visual inspection of the graph was used (ie, if one of the competing models displayed an incongruous narrowing prediction interval [PI]).
      Once the model was selected, prevalence was forecasted. Forecasting was done using a simulation projected to 2030. Prevalence (per 100,000 persons) was forecasted to 2030, and 95% PIs were calculated from the standard deviation derived during forecasting. PIs are probability limits for a forecast that denote a prediction’s accuracy.
      • Box G.E.P.
      • Jenkins G.M.
      • Reinsel G.C.
      • Ljung G.M.
      Time series analysis: forecasting and control.
      Average annual percentage change, with confidence interval (CI), was calculated for each model based on the forecasted prevalence and associated PI using a separate log binomial model on forecasted values only. The total number of people diagnosed with IBD in 2030 and age- and sex-stratified and disease-specific subsets for CD and UC were calculated using forecasted prevalence multiplied by forecasted population values from Statistics Canada.

      Statistics Canada. Table 052-0006 1 components of projected population growth, by projection scenario, Canada, provinces and territories annual (persons x 1,000). Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0520006&&pattern=&stByVal=1&p1=1&p2=9&tabMode=dataTable&csid=. Published 2017. Accessed September 23, 2017.

      Analyses were performed using STATA 14.
      The interactive web-based map was created with ArcGIS Pro 2.3.0 and ArcGIS Online (Environmental Systems Research Institute, Redlands, CA).

       Sensitivity Analyses

      Four sensitivity analyses were conducted. The first analysis used log binomial models instead of ARIMA. The second sensitivity analysis used ARIMA to contrast the total number of people with IBD (ie, total count) with the calculated prevalence (ie, per 100,000).

      Statistics Canada. Estimates of population, by age group and sex for July 1, Canada, provinces and territories annual. Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0510001&tabMode=dataTable&srchLan=-1&p1=-1&p2=9. Published 2017. Accessed June 5, 2017.

      These first 2 sensitivity analyses were done to assess the validity of the results from the primary analysis. The third sensitivity analysis evaluated provinces with validated coding algorithms for their study populations (AB, MB, and ON); this was done to assess the extent of misclassification bias from provinces with non-validated algorithms. The prevalence for Canada was evaluated using a longer period (2002–2013) to assess the effect that a longer period has on these estimates. In this last sensitivity analysis, NS and QC were removed because data were not available for those provinces for the full 12-year period.

      Results

       Provincial Analyses

      The age- and sex-standardized prevalence of IBD across Canada in 2008 ranged from 445 per 100,000 in QC to 870 per 100,000 in NS (Table 1). The forecasted prevalence of IBD in 2018 for each province ranged from 652 per 100,000 (95% PI 619–686) in MB to 1224 per 100,000 (95% PI 1156–1292) in NS (Table 1 and Figure 1). In 2030, the forecasted prevalence of IBD ranged from 819 per 100,000 (95% PI 723–917) in MB to 1657 per 100,000 (95% PI 1531–1783) in NS (Table 1 and Figure 1). Figure 1 presents the annual actual prevalence of IBD for each province followed by the forecasted prevalence, with associated 95% PI. The forecasted prevalence average annual percentage changes ranged from 2.00% (95% CI 1.29–2.61) in MB to 3.89% (95% CI 3.85–3.94) in ON (Table 1 and Figure 1).
      Table 1Actual and Forecasted Prevalence, Including AAPC, Stratified by Province and Age
      IBDCDUC
      Standardized prevalence per 100,000Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)Standardized prevalence per 100,000Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)Standardized prevalence per 100,000Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)
      200820182030200820182030200820182030
      AB529729 (686–771)1048 (882–1214)3.14 (2.12–3.96)282352 (333–372)482 (403–562)2.71 (1.61–3.59)183256 (240–273)376 (312–441)3.33 (2.22–4.21)
      BC515682 (659–705)912 (841–984)2.54 (2.07–2.96)228295 (285–306)390 (357–422)2.41 (1.92–2.85)259348 (338–358)466 (433–499)2.55 (2.14–2.93)
      MB567652 (619–686)819 (723–917)2.00 (1.29–2.61)283316 (299–333)390 (340–439)1.84 (1.07–2.49)285338 (321–356)433 (384–483)2.18 (1.51–2.75)
      NS8701224 (1156–1292)1657 (1531–1783)2.86 (2.55–3.14)412554 (522–587)728 (667–789)2.55 (2.20–2.86)350503 (474–532)692 (634–750)3.03 (2.68–3.34)
      ON507731 (728–735)1156 (1144–1169)3.89 (3.85–3.94)243335 (332–337)500 (493–508)3.40 (3.34–3.47)247363 (360–365)591 (582–600)4.16 (4.09–4.22)
      QC445671 (653–690)940 (905–975)3.25 (3.12–3.36)278427 (415–439)604 (581–627)3.37 (3.24–3.49)168244 (232–256)336 (314–357)3.03 (2.83–3.21)
      SK555636 (606–666)893 (707–1080)2.89 (1.29–4.13)316358 (343–373)491 (399–583)2.69 1.26–3.82)239279 (264–294)403 (313–493)3.13 (1.41–4.43)
      All510725 (716–735)981 (963–999)2.86 (2.80–2.92)263368 (363–373)493 (483–502)2.75 (2.69–2.81)226322 (318–326)436 (428–444)2.87 (2.81–2.93)
      Sex stratification
       Female542768 (758–779)1036 (1017–1055)2.83 (2.77–2.88)296409 (403–415)542 (531–554)2.65 (2.59–2.72)224321 (317–325)437 (429–445)2.92 (2.87–2.98)
       Male477682 (671–692)925 (905–946)2.89 (2.82–2.97)229326 (321–331)441 (432–450)2.87 (2.80–2.94)229323 (319–327)435 (426–443)2.81 (2.75–2.88)
      Age stratification
       Pediatric (<18)
      Younger than 18 years for AB, BC, MB, ON, QC, and SK and <20 years for NS.
      6296 (88–104)159 (133–185)4.32 (3.57–4.93)4058 (52–64)91 (72–109)3.76 (2.81–4.50)1929 (25–34)51 (35–66)4.59 (3.05–5.66)
       Adult (18–59)
      From 18 to 59 years for AB, BC, MB, ON, QC, and SK and 20 to 59 for NS.
      622849 (823–876)1118 (1069–1168)2.59 (2.43–2.74)336453 (430–476)590 (547–633)2.48 (2.24–2.70)262357 (351–364)470 (457–483)2.56 (2.47–2.65)
       Elderly (≥60)646976 (950–1002)1370 (1312–1429)3.24 (3.06–3.40)275427 (416–438)610 (586–633)3.42 (3.27–3.56)340500 (486–513)691 (660–721)3.07 (2.89–3.24)
      AAPC, average annual percentage change.
      a Younger than 18 years for AB, BC, MB, ON, QC, and SK and <20 years for NS.
      b From 18 to 59 years for AB, BC, MB, ON, QC, and SK and 20 to 59 for NS.
      Figure thumbnail gr1
      Figure 1Actual and forecasted prevalence of IBD in Canada by province. Actual prevalence, standardized for age and sex, is denoted by the solid line. Forecasted prevalence—analyzed with an ARIMA model and then forecasted until 2030—is indicated by a dashed line with the PI highlighted in gray. For an interactive map please see https://people.ucalgary.ca/∼ggkaplan/IBDCPREV.html

       National Analyses

      In 2008, the calculated prevalence of IBD in Canada was 510 per 100,000, with CD at 263 per 100,000 and UC at 226 per 100,000 (Table 1 and Figure 2), which equates to 169,564 individuals with IBD. In 2018, the prevalence in Canada was estimated at 725 per 100,000 (95% PI 716–735) for IBD, 368 per 100,000 (95% PI 363–373) for CD, and 322 per 100,000 (95% PI 318–326) for UC; this equates to 267,983 (95% PI 264,579–271,387) individuals with IBD, 135,899 (95% PI 134,065–137,734) with CD, and 118,918 (95% PI 117,424–120,412) with UC (Table 1 and Figure 2).
      Figure thumbnail gr2
      Figure 2Actual and forecasted prevalence of IBD in male and female patients and total patients in Canada. Actual prevalence, standardized for age and sex, is denoted by the solid line and prevalence is calculated through summation of total affected individuals and total population from each province. Forecasted prevalence—analyzed with an ARIMA model and then forecasted until 2030—is indicated by a dashed line with the PI highlighted in gray.
      The average annual percentage change of IBD prevalence in Canada was estimated at 2.86% (95% CI 2.80–2.92; Table 1). The forecasted IBD prevalence in 2030 was 981 per 100,000 (95% PI 963–999), including 493 per 100,000 (95% PI 483–502) with CD and 436 per 100,000 (95% PI 428–444) with UC (Table 1 and Figure 2). The total number of individuals with IBD in 2030 was forecasted at 402,853 (95% PI 395,466–410,240) with IBD, 202,216 (95% PI 198,299–206,133) with CD, and 178,909 (95% PI 175,635–182,184) with UC (Table 1).
      When stratifying the population by sex, the male prevalence of IBD was 477 per 100,000 persons in 2008, increasing to 682 per 100,000 persons (95% PI 671–692) in 2018 and 925 per 100,000 persons (95% PI 905–946) in 2030 (Table 1 and Figure 2). The female prevalence showed similar trends, with a prevalence of 542 per 100,000 persons in 2008, 768 per 100,000 persons (95% PI 758–779) in 2018, and 1036 per 100,000 persons (95% PI 1017–1055) in 2030. These 2030 prevalence estimates equate to approximately 188,326 male persons (95% PI 184,110–192,542) and 214,402 female persons (95% PI 210,487–218,317) with IBD (Table 1 and Figure 2). When stratified by age group, the 2008 prevalence of IBD was 62 per 100,000 in children, 622 per 100,000 in adults, and 646 per 100,000 in the elderly (Table 1 and Figure 3). In 2030, the estimated prevalence of IBD was 159 per 100,000 (95% PI 133–185) in children, 1118 per 100,000 (95% PI 1069–1168) in adults, and 1370 per 100,000 (95% PI 1312–1429) in the elderly (Table 1 and Figure 3). These prevalence estimates equate to approximately 12,647 children (95% PI 10,592–14,702), 238,915 adults (95% PI 228,267–249,563), and 160,736 elderly adults (95% PI 153,898–167,574) living with IBD in 2030 (Table 1 and Figure 3). All estimates for CD and UC by sex or age group are presented in Table 1.
      Figure thumbnail gr3
      Figure 3(A) Actual and forecasted prevalence of pediatric IBD, CD, and UC in Canada. Actual prevalence, standardized for sex and age, is denoted by the solid line. Forecasted prevalence—analyzed with a log binomial model and then forecasted until 2030—is indicated by a dashed line with the PI highlighted in gray. (B) Actual and forecasted prevalence of adult IBD, CD, and UC in Canada. Actual prevalence, standardized for age and sex, is denoted by the solid line. Forecasted prevalence—analyzed with an ARIMA model and then forecasted until 2030—is indicated by a dashed line with the PI highlighted in gray. (C) Actual and forecasted prevalence of elderly IBD, CD, and UC in Canada. Actual prevalence, standardized for age and sex, is denoted by the solid line. Forecasted prevalence—analyzed with an ARIMA model in which the data underwent differencing and then forecasted until 2030—is indicated by a dashed line with the PI highlighted in gray.

       Sensitivity Analyses

      The first sensitivity analysis using log binomial regression yielded a forecasted IBD prevalence in 2018 at 828 per 100,000 (95% PI 816–839), with an increase to 1459 per 100,000 (95% PI 1421–1497) in 2030 (Appendix 2).
      The second sensitivity analysis performed on the total count compared with the calculated prevalence (per 100,000) showed prevalence values that were marginally lower than those found in the primary analysis; this resulted in a forecasted 2030 prevalence of 856 per 100,000 (95% PI 835–876) for all IBD cases, 493 per 100,000 (95% PI 483–502) for CD, and 380 per 100,000 (95% PI 372–388) for UC (see Appendix 2 for further details).
      The third sensitivity analysis, using only provinces with validated algorithms, forecasted a prevalence of IBD at 909 per 100,000 (95% PI 842–976) in 2030 (Appendix 2); this 95% PI overlaps with the 95% PI in the primary analysis, suggesting similar results from the primary and sensitivity analyses (Table 1 and Appendix 2).
      The fourth sensitivity analysis, removing NS and QC from the model to allow for a longer period of analysis, resulted in values similar to those seen in the primary analysis, with an IBD prevalence estimate of 700 per 100,000 (95% PI 680–719) in 2018 and 931 per 100,000 (95% PI 882–980) in 2030 (Appendix 2). An interactive web-based map describing the prevalence of IBD across Canada from 2002 to 2030 is available at this link: https://people.ucalgary.ca/∼ggkaplan/IBDCPREV.html.

      Discussion

      We conducted a nationwide study using historical population-based data from 7 provinces to estimate the current, and forecast the future, prevalence of IBD. In 2008, approximately 0.5% of the Canadian population had IBD; by 2018, prevalence was estimated at 0.7%; and, by 2030, it was forecasted to increase to 1.0%. We estimate that approximately 270,000 Canadians are currently living with IBD and that, by 2030, the Canadian health care systems might be caring for more than 402,000 patients with IBD. Prevalence is increasing in all age groups, but particularly among the elderly because of newly diagnosed seniors and the advancing age of patients with previously diagnosed IBD. Ambulatory clinics in 2030 will be distinctly different from current clinics as health care providers contend with caring for considerably more patients with IBD—including younger newly diagnosed patients and older patients with longer disease duration and comorbidities of advancing age.
      The prevalence of IBD is forecasted to increase steadily by 2.86% per year in Canada. As a lifelong disease without a cure, the compounding prevalence of IBD is due to the disparity between incidence and mortality. Because changes in incidence are inherently captured within our models of changing prevalence (a combination of new cases, current cases, and cases removed from the population), we did not separately analyze incidence. As long as the incidence exceeds the mortality, our clinics will continue to add newly diagnosed patients on the foundation of our previously diagnosed patients every year. However, future studies accounting for life expectancy are necessary to establish the threshold that incidence needs to decrease for prevalence to stabilize. Nonetheless, these data should serve as a clarion call to health care systems to prepare their infrastructure, personnel, and resources to manage the rising burden of IBD.
      With 95% of the Canadian population represented by our study, the findings are generalizable to other westernized nations with similar rates and demographics of IBD. For example, similar to Canada, the prevalence of IBD is increasing steadily in the United States.
      • Kaplan G.G.
      • Ng S.C.
      Globalisation of inflammatory bowel disease: perspectives from the evolution of inflammatory bowel disease in the UK and China.
      • Kappelman M.D.
      • Moore K.R.
      • Allen J.K.
      • Cook S.F.
      Recent trends in the prevalence of Crohn’s disease and ulcerative colitis in a commercially insured US population.
      A series of studies documenting this trend in Olmsted County, Minnesota was published using data ranging from 1940 to 2010.
      • Loftus C.G.
      • Loftus Jr., E.V.
      • Harmsen W.S.
      • et al.
      Update on the incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota, 1940–2000.
      • Shivashankar R.
      • Tremaine W.J.
      • Harmsen W.S.
      • Loftus Jr., E.V.
      Incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota from 1970 through 2010.
      In 2001, the prevalence of IBD in Olmsted County was 388 per 100,000 and increased to 533 per 100,000 in 2010; this equates to an estimated 1.64 million individuals living with IBD as of 2010 in the United States.
      • Loftus C.G.
      • Loftus Jr., E.V.
      • Harmsen W.S.
      • et al.
      Update on the incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota, 1940–2000.
      • Shivashankar R.
      • Tremaine W.J.
      • Harmsen W.S.
      • Loftus Jr., E.V.
      Incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota from 1970 through 2010.

      United States Census Bureau. Decennial census datasets. Available at: https://www.census.gov/programs-surveys/decennial-census/data/datasets.2010.html. Published 2018. Accessed May 11, 2018.

      If the prevalence increases to the same level as in Canada—0.98% in 2030—more than 3.48 million individuals could have IBD in the United States, more than double the 2010 estimate.

      United States Census Bureau. Bureau USC. 2017 National population projections tables. Available at: https://www.census.gov/data/tables/2017/demo/popproj/2017-summary-tables.html. Published 2017. Accessed May 11, 2018.

      Moreover, similar trends of rising prevalence have been observed in Europe and Australia.
      • Kaplan G.G.
      • Ng S.C.
      Globalisation of inflammatory bowel disease: perspectives from the evolution of inflammatory bowel disease in the UK and China.
      Thus, a strategy to address the compounding prevalence of IBD in the Western world is necessary. Although our data are less applicable outside the Western world, where prevalence of IBD remains low, the rapidly increasing incidence of IBD in newly industrialized countries in Asia and Latin America suggests that these countries might begin to experience compounding prevalence analogous to the Western world in a generation from now.
      • Kaplan G.G.
      • Ng S.C.
      Globalisation of inflammatory bowel disease: perspectives from the evolution of inflammatory bowel disease in the UK and China.
      • Rocchi A.
      • Benchimol E.I.
      • Bernstein C.N.
      • et al.
      Inflammatory bowel disease: a Canadian burden of illness review.
      Although the exact future prevalence of IBD is unknown, it is apparent that the prevalence will continue to increase—so, how do health care systems adapt? By bringing awareness of this impending burden to the public and physicians alike, we can increase community engagement and bring this disease to the forefront of policy makers’ agendas. This awareness could lead to an influx of funding for laboratory and epidemiologic research, which in turn will aid in preventing disease by furthering the knowledge base and thus working to lower the incidence of disease. If we cannot prevent new cases of IBD, then the forecasted increases in prevalence will lead to increased overall costs to health care systems, especially with medications, some of which can exceed tens of thousands of dollars per year per individual. Identifying the current drivers of cost and projecting the costs into the future are paramount to being able to define the future burden of disease, evaluate the future impact on health care systems, and ensure access and affordability. Further, developing alternative care pathways (eg, nurse practitioners) will help to ensure that those with IBD still receive the necessary care without health care systems being overwhelmed.
      This study forecasts the future prevalence of IBD, but it is not without limitations. Heterogeneity in prevalence was observed between provinces. NS had the highest prevalence of IBD. Because NS is the only Atlantic province in the dataset, we could not assess whether the high prevalence value was representative of the east coast of Canada or unique to NS. Heterogeneity between provinces can be explained by inherent differences between populations, such as demography, genetic penetrance, and environmental exposures associated with IBD (eg, diet).
      • Rocchi A.
      • Benchimol E.I.
      • Bernstein C.N.
      • et al.
      Inflammatory bowel disease: a Canadian burden of illness review.
      The models do not directly account for future variations in environmental risk factors (eg, breastfeeding or smoking) that can influence incidence, because these data were not available at the individual level for this cohort.
      • Molodecky N.A.
      • Panaccione R.
      • Ghosh S.
      • et al.
      Alberta Inflammatory Bowel Disease Consortium. Challenges associated with identifying the environmental determinants of the inflammatory bowel diseases.
      Further, because prevalence was calculated using administrative health care databases, possible misclassification errors in the diagnosis of IBD can occur.
      • Molodecky N.A.
      • Panaccione R.
      • Ghosh S.
      • et al.
      Alberta Inflammatory Bowel Disease Consortium. Challenges associated with identifying the environmental determinants of the inflammatory bowel diseases.
      • Sorensen H.T.
      • Sabroe S.
      • Olsen J.
      A framework for evaluation of secondary data sources for epidemiological research.
      • Rezaie A.
      • Quan H.
      • Fedorak R.N.
      • et al.
      Development and validation of an administrative case definition for inflammatory bowel diseases.
      • Oleckno W.A.
      Epidemiology: concepts and methods.
      To overcome this, AB, MB, and ON use validated algorithms, which ensure accuracy of the data based on sensitivity, specificity, and predictive values of diagnostic codes. The other provinces have applied these algorithms without cross-referencing accuracy with chart reviews. However, our sensitivity analysis—evaluating only the provinces with validated algorithms—produced prevalence estimates consistent with our main analyses.
      • Molodecky N.A.
      • Panaccione R.
      • Ghosh S.
      • et al.
      Alberta Inflammatory Bowel Disease Consortium. Challenges associated with identifying the environmental determinants of the inflammatory bowel diseases.
      • Rezaie A.
      • Quan H.
      • Fedorak R.N.
      • et al.
      Development and validation of an administrative case definition for inflammatory bowel diseases.
      Also, our national estimates of IBD include all patients with IBD; however, 5.6% of patients have unclassified IBD. These patients have IBD, but the algorithm used to distinguish CD from UC cannot differentiate the specific subtype. In consequence, the CD and UC stratified analyses do not account for these individuals. In addition, provinces report different timeframes based on available data: MB contains the longest period (1990–2013) and QC contains the shortest (2001–2008). Thus, our primary national model was restricted to 2002–2008 that included prevalence data from all 7 provinces. However, a sensitivity analysis of 5 provinces with prevalence data spanning 2002–2013 yielded similar estimates on forecasted prevalence (931 per 100,000 in 2030) compared with our primary national model (981 per 100,000). Further, there is an inherent uncertainty associated with forecast modeling because the methodology forecasts future values that are yet unknown based on an assumption that historical trends (eg, incidence rates) will be similar in the future. To address this, PIs are created that acknowledge this uncertainty and give a range of possible values for the true prevalence. The use of the best fitting models also mitigates this limitation by ensuring that forecasting models are true to the current data. Moreover, sensitivity analyses using alternate forecasting approaches (ie, log binomial modeling) yield similar estimates.
      Forecasting the number of people with IBD alone is insufficient in defining the overall burden of IBD. We anticipate that the future landscape of IBD will evolve: increased use of biologics and new therapies in the pipeline, decreasing rates of hospitalization and surgery, and unpredictability of a landmark discovery, including a potential cure.
      • Amiot A.
      • Peyrin-Biroulet L.
      Current, new and future biological agents on the horizon for the treatment of inflammatory bowel diseases.
      • Frolkis A.D.
      • Dykeman J.
      • Negron M.E.
      • et al.
      Risk of surgery for inflammatory bowel diseases has decreased over time: a systematic review and meta-analysis of population-based studies.
      Future studies are necessary to integrate rising prevalence with other clinical factors that influence the burden of IBD to the health care system.
      Overall, the future prevalence of IBD will lead to an increased stress on health care systems. We are at a time when new policies and innovations in the delivery of IBD care can help ensure that the nearly 4 million patients with IBD in North America (by 2030) will continue to receive the necessary care in the future. If health care systems fail to adjust for the impending burden of IBD, then they will be overwhelmed, and patients might not receive the care they need.

      Acknowledgments

      Disclaimer: This study is based in part on de-identified data provided by Alberta Health Services, Régie de l’assurance maladie du Québec, Ministry of Health of British Columbia and Population Data BC, Saskatchewan Ministry of Health, Health Data Nova Scotia of Dalhousie University, and Manitoba Health. The interpretation and conclusions contained herein are those of the researchers and do not necessarily represent the views of the Government of Alberta, Government of Quebec, Government of British Columbia, Government of Saskatchewan or the Ministry of Health, Health Data Nova Scotia or the Department of Health and Wellness, or the Government of Manitoba. Neither the Government of Alberta nor Alberta Health Services expressed any opinion in relation to this study. This study was supported by the Institute for Clinical Evaluative Sciences, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred. The interpretation and conclusions contained herein do not necessarily represent those of the Government of Saskatchewan or the Ministry of Health. Although this research and health service assessment analysis is based on data obtained from the Nova Scotia Department of Health and Wellness, the observations and opinions expressed are those of the authors and do not represent those of Health Data Nova Scotia or the Department of Health and Wellness.
      An interactive web-based map describing the prevalence of IBD across Canada from 2002 to 2030 is available at this link: https://people.ucalgary.ca/∼ggkaplan/IBDCPREV.html
      Author contributions: Stephanie Coward contributed to study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; and statistical analysis. Fiona Clement, Charles N. Bernstein, Mathew W. Carroll, Glen Hazlewood, Kevan Jacobson, M. Ellen Kuenzig, Des Leddin, Kerry A. McBrien, Sanjay K. Murthy, Remo Panaccione, Ali Rezaie, and Harminder Singh contributed to study design and critical revision of the manuscript for important intellectual content. Eric I. Benchimol, J. Antonio Avina-Zubieta, Alain Bitton, Susan Jelinski, Jennifer L. Jones, Geoffrey C. Nguyen, Anthony R. Otley, Greg Rosenfeld, Juan Nicolás Peña-Sánchez, and Laura E. Targownik contributed to study design, data acquisition, critical revision of the manuscript for important intellectual content, and obtained funding. Rob Deardon and Ali Rezaie contributed to study design, interpretation of data, and critical revision of the manuscript for important intellectual content. Gilaad G. Kaplan contributed to study concept and design; acquisition of data; interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; obtained funding; administrative, technical, or material support; and study supervision.

      Supplementary Material

      Appendix 1Province-Specific Overview
      Prevalent cohort availabilityABBCMBNSONQCSK
      2002–20151997–20141990–20131996–20091999–20142001–20081998–2016
      Time measureFiscal year (April 1–March 31)Calendar year (January 1–December 31)Mid-year (July 1)Calendar year (January 1–December 31)Fiscal year (April 1–March 31)Mid-year (July 1)Fiscal year (April 1–March 31)
      Diseases classifiedCD, UC, IBD-U, IBDCD, UC, IBD-U, IBDCD, UC, IBDCD, UC, IBD-U, IBDCD, UC, IBD-U, IBDCD, UC, IBDCD, UC, IBD
      Identification algorithm validation referenceRezaie et al, 2012
      • Rezaie A.
      • Quan H.
      • Fedorak R.N.
      • et al.
      Development and validation of an administrative case definition for inflammatory bowel diseases.
      Rezaie et al, 2012
      • Rezaie A.
      • Quan H.
      • Fedorak R.N.
      • et al.
      Development and validation of an administrative case definition for inflammatory bowel diseases.
      Bernstein et al, 1999
      • Bernstein C.N.
      • Blanchard J.F.
      • Rawsthorne P.
      • Wajda A.
      Epidemiology of Crohn’s disease and ulcerative colitis in a central Canadian province: a population-based study.
      N/AAdults: Benchimol et al, 2014
      • Benchimol E.I.
      • Guttmann A.
      • Mack D.R.
      • et al.
      Validation of international algorithms to identify adults with inflammatory bowel disease in health administrative data from Ontario, Canada.
      ; pediatrics: Benchimol et al, 2009
      • Benchimol E.I.
      • Guttmann A.
      • Griffiths A.M.
      • et al.
      Increasing incidence of paediatric inflammatory bowel disease in Ontario, Canada: evidence from health administrative data.
      Rezaie et al, 2012
      • Rezaie A.
      • Quan H.
      • Fedorak R.N.
      • et al.
      Development and validation of an administrative case definition for inflammatory bowel diseases.
      Bernstein et al, 1999
      • Bernstein C.N.
      • Blanchard J.F.
      • Rawsthorne P.
      • Wajda A.
      Epidemiology of Crohn’s disease and ulcerative colitis in a central Canadian province: a population-based study.
      Identification algorithm usedAB (≥2 hospitalizations or 4 physician claims or 2 medical contacts in 2 y)AB (≥2 hospitalizations or 4 physician claims or 2 medical contacts in 2 y)MB (5 physician contacts of any combination of outpatient contacts or hospitalizations using Outpatient Physician Database or Hospitalization Database)AB (≥2 hospitalizations or 4 physician claims or 2 medical contacts in 2 y)Adults 18–64 y: 5 physician contacts or hospitalizations within 4 y; adults ≥65 y: pharmacy claim for IBD medication + 5 physician contacts or hospitalizations within 4 y; children <18 y: if scoped: 4 OHIP or 2 CIHI-DAD within 3 y; if not scoped: 7 OHIP or 2 CIHI-DAD within 3 yAB (≥2 hospitalizations or 4 physician claims or 2 medical contacts in 2 y)MB (≥5 physician contacts or CIHI-DAD records within 2 y of health coverage, and ≥3 separate contacts with <2 y of coverage)
      Validity of identification algorithmSensitivity 83.4%, positive predictive value 97.4%N/ASensitivity 74.4%–89.2%, specificity 89.8%–93.7%N/APediatrics (<18 y): sensitivity 86.9%–91.1%, positive predictive value 57.7%–75.2%; adults (18–64 y): sensitivity 76.8%–92.3%, positive predictive value 81.4%; elderly (>64 y): sensitivity 59.3%–78.3%, positive predictive value 71.1%N/AN/A
      Physician billing (associated with 1 diagnostic code)Alberta Health Physician ClaimsPopulation Data BC captures all provincially funded health care services data since 1990, including all outpatient medical visits,

      British Columbia Ministry of Health. Population Data BC. Medical services plan (MSP) payment information file. Data extract. MOH. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      hospital admissions and discharges,

      British Columbia Ministry of Health. Population Data BC. Discharge abstract database (hospital separations). Data extract. MOH. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      interventions,

      British Columbia Ministry of Health. Population Data BC. Medical services plan (MSP) payment information file. Data extract. MOH. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      investigations,

      British Columbia Ministry of Health. Population Data BC. Medical services plan (MSP) payment information file. Data extract. MOH. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      demographic data,

      British Columbia Ministry of Health. Population Data BC. Consolidation file (MSP registration & premium billing). Data extract. MOH. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      cancer registry,

      BC Cancer Agency. Population Data BC. BC Cancer Agency registry data 2014. Available at: http://www.popdata.bc.ca/data. Published 2014. Accessed May 15, 2018.

      and vital statistics.

      BC Vital Statistics Agency. Population Data BC. Vital statistics deaths. Data extract BC vital statistics agency. Available at: http://www.popdata.bc.ca/data. Published 2012. Accessed May 15, 2018.

      Furthermore, Population Data BC encompasses the comprehensive prescription drug database PharmaNet

      BC Ministry of Health. PharmaNet. Data extract. Data stewardship committee. Available at: http://www.popdata.bc.ca/data. Published 2013. Accessed May 15, 2018.

      Manitoba Health Physicians Claims databaseMSIOHIPRAMQPhysician Services Claims File: Medical Services Branch (MSB)
      Hospitalization (associated with 20 diagnostic codes)CIHI-DADCIHI-DADCIHI-DADCIHI-DADMedEchoHospital Discharge Abstract Database (CIHI-DAD)
      Ambulatory care (including ED visits)NACRSManitoba Health Physicians Claims databaseOnly day surgery and ED visitsOHIP (1991–2016); NACRS (2002 onward)MedEchoMSB and ED visits from NACRS (CIHI NACRS, ED data)
      Basic demographic information (date of birth, eligibility, date of death)Alberta Health RegistryManitoba Health population registryNova Scotia Vital Statistics, Insured Patient RegistryRPDBRAMQPHRS
      Sample of previously published articles using algorithmFrolkis et al, 2014
      • Frolkis A.
      • Kaplan G.G.
      • Patel A.B.
      • et al.
      Postoperative complications and emergent readmission in children and adults with inflammatory bowel disease who undergo intestinal resection: a population-based study.
      ; Kuenzig et al, 2017
      • Kuenzig M.E.
      • Barnabe C.
      • Seow C.H.
      • et al.
      Asthma is associated with subsequent development of inflammatory bowel disease: a population-based case-control study.
      N/ABenchimol et al, 2017
      • Benchimol E.I.
      • Bernstein C.N.
      • Bitton A.
      • et al.
      Trends in epidemiology of pediatric inflammatory bowel disease in Canada: distributed network analysis of multiple population-based provincial health administrative databases.
      ; Melesse et al, 2015
      • Melesse D.Y.
      • Targownik L.E.
      • Singh H.
      • et al.
      Patterns and predictors of long-term nonuse of medical therapy among persons with inflammatory bowel disease.
      and 2017
      • Melesse D.Y.
      • Lix L.M.
      • Nugent Z.
      • et al.
      Estimates of disease course in inflammatory bowel disease using administrative data: a population-level study.
      ; Shaw et al, 2013,
      • Shaw S.Y.
      • Blanchard J.F.
      • Bernstein C.N.
      Association between early childhood otitis media and pediatric inflammatory bowel disease: an exploratory population-based analysis.
      2014,
      • Shaw S.Y.
      • Nugent Z.
      • Targownik L.E.
      • et al.
      Association between spring season of birth and Crohn’s disease.
      and 2015
      • Shaw S.Y.
      • Blanchard J.F.
      • Bernstein C.N.
      Early childhood measles vaccinations are not associated with paediatric IBD: a population-based analysis.
      Leddin et al, 2014
      • Leddin D.
      • Tamim H.
      • Levy A.R.
      Decreasing incidence of inflammatory bowel disease in Eastern Canada: a population database study.
      Benchimol et al, 2014,
      • Benchimol E.I.
      • Manuel D.G.
      • Guttmann A.
      • et al.
      Changing age demographics of inflammatory bowel disease in Ontario, Canada: a population-based cohort study of epidemiology trends.
      2014,
      • Benchimol E.I.
      • Mack D.R.
      • Nguyen G.C.
      • et al.
      Incidence, outcomes, and health services burden of very early onset inflammatory bowel disease.
      and 2015
      • Benchimol E.I.
      • Mack D.R.
      • Guttmann A.
      • et al.
      Inflammatory bowel disease in immigrants to Canada and their children: a population-based cohort study.
      ; Bollegala et al, 2017
      • Bollegala N.
      • Benchimol E.I.
      • Griffiths A.M.
      • et al.
      Characterizing the posttransfer period among patients with pediatric onset IBD: the impact of academic versus community adult care on emergent health resource utilization.
      ; Nguyen et al, 2015
      • Nguyen G.C.
      • Sheng L.
      • Benchimol E.I.
      Health care utilization in elderly onset inflammatory bowel disease: a population-based study.
      Bitton et al, 2014
      • Bitton A.
      • Vutcovici M.
      • Patenaude V.
      • et al.
      Epidemiology of inflammatory bowel disease in Quebec: recent trends.
      Pena-Sanchez et al, 2017
      • Pena-Sanchez J.N.
      • Lix L.M.
      • Teare G.F.
      • et al.
      Impact of an integrated model of care on outcomes of patients with inflammatory bowel diseases: evidence from a population-based study.
      Cases of IBD in 2008 (real)
      Numbers presented might vary from actual numbers in the original administrative data: suppressed cells occurred in data provided by provinces, which altered the final values contained in the analysis compared with the original.
      18,26922,9376516810464,96335,0055357
      Cases of IBD in 2008 (standardized)
      Numbers were standardized by age and sex based on the Canadian population.27
      19,03922,4206795814365,27834,5675643
      Population size in 20083,595,7554,499,1391,197,774935,86512,882,6257,761,5041,017,346
      CIHI-DAD, Canadian Institute for Health Information–Discharge Abstract Database; ED, emergency department; IBD-U, inflammatory bowel disease–unclassified; MSB, Medical Services Billing; MSI, Medical Services Incorporated; N/A, not applicable; NACRS, National Ambulatory Care Reporting System; OHIP, Ontario Health Insurance Plan; PHRS, Person Health Registration System; RAMQ, Régie de l’assurance maladie du Québec; RPDB, Registered Persons Database.
      a Numbers presented might vary from actual numbers in the original administrative data: suppressed cells occurred in data provided by provinces, which altered the final values contained in the analysis compared with the original.
      b Numbers were standardized by age and sex based on the Canadian population.27
      Appendix 2Sensitivity Analyses of Forecasted Prevalence and AAPC
      IBDCDUC
      Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)Forecasted prevalence per 100,000 (95% PI)Forecasted AAPC, % (95% CI)
      201820302018203020182030
      All log binomial regression828 (816–839)1459 (1421–1497)4.85 (4.74–4.96)414 (406–422)706 (680–732)4.55 (4.40–4.70)368 (360–375)652 (626–678)4.89 (4.72–5.05)
      Using population values684 (673–694)856 (835–876)2.23 (2.13–2.32)368 (363–373)493 (483–502)2.75 (2.69–2.81)303 (300–307)380 (372–388)2.23 (2.15–2.32)
      Canada (validated algorithms)681 (661–701)909 (842–976)2.52 (2.09–2.91)319 (307–330)412 (376–449)2.25 (1.71–2.74)309 (289–328)413 (357–469)2.55 (1.74–3.23)
      Canada (without NS or QC: 2002–2013)700 (680–719)931 (882–980)2.52 (2.25–2.77)327 (317–336)425 (400–450)2.33 (2.02–2.61)328 (316–340)438 (409–468)2.57 (2.24–2.88)
      AAPC, average annual percentage change.

      References

        • Molodecky N.A.
        • Soon I.S.
        • Rabi D.M.
        • et al.
        Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review.
        Gastroenterology. 2012; 142: 46-54
        • Ng S.C.
        • Shi H.Y.
        • Hamidi N.
        • et al.
        Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies.
        Lancet. 2018; 390: 2769-2778
        • Kaplan G.G.
        • Ng S.C.
        Globalisation of inflammatory bowel disease: perspectives from the evolution of inflammatory bowel disease in the UK and China.
        Lancet Gastroenterol Hepatol. 2016; 1: 307-316
        • Benchimol E.I.
        • Fortinsky K.J.
        • Gozdyra P.
        • et al.
        Epidemiology of pediatric inflammatory bowel disease: a systematic review of international trends.
        Inflamm Bowel Dis. 2011; 17: 423-439
        • Rocchi A.
        • Benchimol E.I.
        • Bernstein C.N.
        • et al.
        Inflammatory bowel disease: a Canadian burden of illness review.
        Can J Gastroenterol. 2012; 26: 811-817
        • Kaplan G.G.
        The global burden of IBD: from 2015 to 2025.
        Nat Rev Gastroenterol Hepatol. 2015; 12: 720-727
      1. Statistics Canada. Estimates of population, by age group and sex for July 1, Canada, provinces and territories annual. Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0510001&tabMode=dataTable&srchLan=-1&p1=-1&p2=9. Published 2017. Accessed June 5, 2017.

        • Bujang M.A.
        • Adnan T.H.
        • Hashim N.H.
        • et al.
        Forecasting the incidence and prevalence of patients with end-stage renal disease in Malaysia up to the year 2040.
        Int J Nephrol. 2017; 2017: 2735296
      2. Adhikari R, Agrawal RK. An introductory study on time series modeling and forecasting. Available at; https://arxiv.org/ftp/arxiv/papers/1302/1302.6613.pdf. Published 2013. Accessed January 22, 2018.

        • Box G.E.P.
        • Jenkins G.M.
        • Reinsel G.C.
        • Ljung G.M.
        Time series analysis: forecasting and control.
        5th ed. John Wiley & Sons, Hoboken, NJ2015
      3. Katchova A. Econometrics Academy. Time series ARIMA models. Available at: https://sites.google.com/site/econometricsacademy/econometrics-models/time-series-arima-models. Published 2015. Accessed January 22, 2018.

      4. Statistics Canada. Table 052-0006 1 components of projected population growth, by projection scenario, Canada, provinces and territories annual (persons x 1,000). Available at: http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=0520006&&pattern=&stByVal=1&p1=1&p2=9&tabMode=dataTable&csid=. Published 2017. Accessed September 23, 2017.

      5. Stata version 14 [computer program]. StataCorp, College Station, TX2015
        • Kappelman M.D.
        • Moore K.R.
        • Allen J.K.
        • Cook S.F.
        Recent trends in the prevalence of Crohn’s disease and ulcerative colitis in a commercially insured US population.
        Dig Dis Sci. 2013; 58: 519-525
        • Loftus C.G.
        • Loftus Jr., E.V.
        • Harmsen W.S.
        • et al.
        Update on the incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota, 1940–2000.
        Inflamm Bowel Dis. 2007; 13: 254-261
        • Shivashankar R.
        • Tremaine W.J.
        • Harmsen W.S.
        • Loftus Jr., E.V.
        Incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota from 1970 through 2010.
        Clin Gastroenterol Hepatol. 2017; 15: 857-863
      6. United States Census Bureau. Decennial census datasets. Available at: https://www.census.gov/programs-surveys/decennial-census/data/datasets.2010.html. Published 2018. Accessed May 11, 2018.

      7. United States Census Bureau. Bureau USC. 2017 National population projections tables. Available at: https://www.census.gov/data/tables/2017/demo/popproj/2017-summary-tables.html. Published 2017. Accessed May 11, 2018.

        • Molodecky N.A.
        • Panaccione R.
        • Ghosh S.
        • et al.
        Alberta Inflammatory Bowel Disease Consortium. Challenges associated with identifying the environmental determinants of the inflammatory bowel diseases.
        Inflamm Bowel Dis. 2011; 17: 1792-1799
        • Sorensen H.T.
        • Sabroe S.
        • Olsen J.
        A framework for evaluation of secondary data sources for epidemiological research.
        Int J Epidemiol. 1996; 25: 435-442
        • Rezaie A.
        • Quan H.
        • Fedorak R.N.
        • et al.
        Development and validation of an administrative case definition for inflammatory bowel diseases.
        Can J Gastroenterol. 2012; 26: 711-717
        • Oleckno W.A.
        Epidemiology: concepts and methods.
        Waveland Press, Long Grove, IL2008
        • Amiot A.
        • Peyrin-Biroulet L.
        Current, new and future biological agents on the horizon for the treatment of inflammatory bowel diseases.
        Therap Adv Gastroenterol. 2015; 8: 66-82
        • Frolkis A.D.
        • Dykeman J.
        • Negron M.E.
        • et al.
        Risk of surgery for inflammatory bowel diseases has decreased over time: a systematic review and meta-analysis of population-based studies.
        Gastroenterology. 2013; 145: 996-1006