The healthy immigrant effect in Canada: a longitudinal perspective using National Population Health Surveys Edward Ng (a), Russell Wilkins, (a, b) and.

The healthy immigrant effect in Canada: a longitudinal perspective using National Population Health Surveys

Edward Ng (a), Russell Wilkins, (a, b) and Jean-Marie Berthelot (a, c)

(a)Health Analysis and Measurement Group, Statistics Canada. (b)Department of Epidemiology and Community Medicine, University of Ottawa.

(c) Department of Epidemiology and Biostatistics, McGill University.

Paper for presentation at the 2004 Annual Meetings of the Canadian Population Society

Winnipeg, Manitoba, 3-5 June 2004

• Immigrants have played and will continue to play important roles in Canada.– In 2001, 5.4 million (18%) of Canada’s population

were first generation foreign-born immigrants, the highest percentage in 70 years.

– This included 1.8 million immigrants who arrived in the period 1991- 2001.

• While potential immigrants are screened in terms of health, relatively little is done to monitor immigrants’ health after entry.

Introduction

• Previous studies in Canada and elsewhere generally point to a strong ‘healthy immigrant effect’ which diminishes over time. – Immigrants, especially recent immigrants, are usually

healthier. However, as time after immigration elapses, their health status tends to converge towards that of the host population.

• However, an important limitation of many earlier studies is the cross-sectional nature of their data.

Healthy immigrant effect

Age-adjusted prevalence of long-term disability by immigrant status and duration of residence

12.8

20.1

17.8

22

21.1

0 5 10 15 20 25

Recent imm.

Long-term Imm.

All immigrants

Canadian-born

Total

%

Source: Chen et al. 1996

Purpose of the study

• To better understand the duration-related changes in immigrants’ health over time in the host country vis-à-vis that of the Canadian-born.

• We used first four waves of the National Population Health Survey (NPHS), a longitudinal panel survey conducted bi-annually from 1994 to 2000.

• There were about 14,100 respondents aged 18 and over from the 1994/95 cycle. – immigrants (15%)

Methods & materials

Survival analysis of health transitions

• Proportional hazards modelling was used as it can handle censoring. – Censoring comes in many forms and occurs for many

different reasons. Right censoring is when observation is terminated before the event occurs.

• Models were fitted to estimate the relative risks of transition for different immigrant status groups with respect to 3 outcomes: (a) self-rated health, (b) health-related behaviour, and (c) health care utilization. – In all analyses, NPHS longitudinal normalized weights

were used.

Explanatory variables• Immigrants’ region of origin was defined by place

of birth and immigration status: – “European” included those born in Europe, the USA,

Oceania and Mexico.– “Non-European” included all other countries of birth.

• Immigrants’ duration of residence was defined as the number of years since immigrating to Canada (as of 1994/95):

1994/951984-85

RecentLong-term

2000/2001

(a) Self-rated health status (“In general, how would you describe your health?”)– Transition from good (good, very good or excellent) to poor

(fair or poor) health

(b) Smoking (Type of smoker) – Transition from not daily smoker to daily smoker

(c) Number of doctors’ consultations in the past year– Transition from < 6 to 6 +

Outcome studied

Estimated immigrant population• In 1994/95, there were 4.3 million immigrants,

representing 20% of the 21 million household population aged 18 and over in Canada.

0

0.5

1

1.5

2

2.5

3

3.5

Long-Term Recent

Mil

lio

ns

Non-European

European

• European immigrants accounted for 58% of these immigrants (some 2.5 million).• Only 25% of the recent immigrants were European (about 320,000).

Chart 1 Sample attrition by immigrant source*duration, NPHS 1994/95 and 2000/01

33

28

31

34

29

30

30

22

23

0 5 10 15 20 25 30 35 40

Recent

LT

Non-European

Recent

LT

European

All Immigrants

Canadian born

Total

%

Results

Chart 2 Adjusted relative risk of change away from self-rated good health, by immigrant source/duration among initially healthy

population, 1994/95-2000/01†

0 0.5 1 1.5 2 2.5

Canadian

*

* Statistically significant at .05 level

*

Recent Non-European

Long-term European

Recent European

Long-tern Non-European

† Controlled for age, sex, low income status, education and smoking behaviour in 1994/95

Chart 3 Adjusted relative risk of becoming a daily smoker, by immigrant source/duration among initially

healthy population, 1994/95-2000/01 †

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Canadian

*

* Statistically significant at .05 level

*

Recent Non-European

Long-term European

Recent European

Long-tern Non-European

† Controlled for age, sex, low income status and. education in 1994/95

Chart 4 Adjusted relative risk of change toward 6 or more doctor’s consultation, by immigrant source/duration among initially healthy

population, 1994/95-2000/01 †

0 0.5 1 1.5 2 2.5

Canadian

*

* Statistically significant at .05 level

Recent Non-European

Long-term European

Recent European

Long-tern Non-European

† Controlled for age, sex, low income status and. education in 1994/95

• Using the first 4 cycles of NPHS, we found that , over time, non-European immigrants (especially those recent immigrants) were significantly more likely to move away from self-rated good health.

– This is not true for the European immigrants.

(a) Self-rated health

Discussion

• This study also found that becoming a daily smoker is probably not likely to be the reason for non-European immigrants’ worsening self-rated health, as these immigrants had low relative risks of adopting daily smoking behaviour.

(b) Health-related behaviour

• Previous cross-sectional studies on the level of health care utilization found that immigrants tend to under-utilize preventive and mental health services, and do not over-utilize health care services as a whole.

• Our longitudinal analysis provides a dynamic view that recent non-European immigrants had significantly higher relative risks of having frequent consultations with doctors, compared to the Canadian born population.

– This is consistent with the findings that they were also significantly more likely to report poorer health over time.

(c) Health care utilization

Concluding remarks

• Due to the heterogeneity of the immigrant populations, the grouping the immigrants into European and non-European sub-groups can at best be a crude way to capture the cultural differences underlying the phenomena studied. – dual pathway of health transition among immigrants?

• It would also be desirable to compare health for refugees and other classes of immigrants. However, refugee status was also not collected in the NPHS data.

• While the use of survival model attempts to minimize the loss of sample due to attrition, it is possible that the convergence of health can be partly due to the higher attrition of healthier immigrants and less healthy Canadian-born population.

• This is research in progress, and will be extended to the 5th wave of NPHS. Our next step in the analysis will be to explore – other health status dimensions (such as disability) – other possible explanations for the change in health risk

factor (leisure activity) and – other health care utilization (such as hospitalization).

Contact informationEdward Ng

Health Analysis and Measurement Group

Statistics Canada, RHC-24R

Ottawa, ON, Canada K1A 0T6

Tel.: 613-951-5308

Fax: 613-951-5939

E-mail: [email protected]