Relationship between poverty and health among adolescents

Relationship between poverty and health among adolescents – Statistical Data Included

Thomas J. Abernathy

Despite a commitment by the Canadian government to eliminate child poverty, the rate of poverty among children and adolescents remains at nearly 20% (Centre for International Statistics, 1994, 1995). The individual and societal implications of this situation are of growing concern. The Canadian Council on Social Development, for instance, has noted that poverty often is accompanied by a higher rate of infant mortality, childhood health problems, psychosocial disorders, and school dropout rates.

Although the association between low socioeconomic status and health is well established (Frank & Mustard, 1994), the mechanisms by which income influences health are less clear. Evans and Stoddart (1994) suggest that health and disease are determined by interactions between influences in the social environment, physical environment, access to health care, and individual behavioral and biological responses. Their model expresses the interrelationship between the determinants of health and measures of health, as well as among the determinants themselves. This study uses the Evans-Stoddart model as a framework to explore data from the 1994 National Population Health Survey regarding the pathways by which income adequacy might influence the health and well-being of young Canadians.

METHOD

The 1994 National Population Health Survey (NPHS) collected information from households across Canada, excluding populations residing on Indian reserves, Canadian military bases, and in remote regions of Quebec and Ontario (Statistics Canada, 1995). Limited information was gathered on all household members, followed by a more detailed interview with a randomly selected household member, aged 12 years or above, which examined health status, health service utilization, behavioral risk factors, psychosocial factors, demographics, and socio-economic status.

A stratified two-stage sampling design, including mechanisms to ensure equal representation of members of larger households, was employed. The overall household response rate was 88.7%; the selected person response rate was 92.5%. A total of 637 youths aged 12 to 14 and 1,210 aged 15 to 19 across Canada were directly interviewed. Partial nonresponse occurred for several individual variables. Information on household income was unavailable for 88 respondents, reducing the sample size for this analysis to 1,759.

Wherever possible, several variables were chosen to represent the determinant category within the model. In some cases, however, the nature of the data collected was such that certain categories contain few measures. As a consequence, the indicators used cannot be considered exhaustive, but rather represent certain components of the determinants.

Variable Definition

Income adequacy. Household income was recoded and grouped into one of five fixed income categories adjusted for family size. This measure is based on all sources of income, including government transfers. It does not define or classify poverty per se, but acts as a crude relative measure of income adequacy based on predetermined cutoff points.

Social and physical environment. Social environment was represented by living arrangement (children living with a single parent compared with all other family types), urban or rural dwelling, and social support index (a five-point scale ranging from low to high support). The presence of a smoker in the house was the only available indicator of physical environment.

Individual response. Behavioral responses included the subject’s smoking, alcohol use, and physical activity. Self-esteem, mastery, and distress scales were used to measure psychological responses.

Health and function. Health and function were conceptualized as the individual’s perception of his or her own as well as its impact on day-to-day functioning. These concepts were measured via self-assessed health (ranging across five categories from poor to excellent), number of disability days in the last two weeks, and whether the individual had experienced an activity-limiting injury in the last 12 months.

Disease. Disease, on the other hand, was viewed as a medical concept and defined by the presence of one or more chronic conditions diagnosed by a health professional (from a list of 17 specific conditions and a more general question asking about “any other long-term condition”). Mental illness was represented by depression, measured on a scale with higher values assigned to increasing numbers of symptoms typically associated with depression as experienced over the last year.

Health care. Health care included consultations with health professionals, physician visits, and acute hospitalizations during the last 12 months. Use of mental health services included the number of social worker/counsellor and psychologist consultations in the last 12 months. Since differential access to health care services may explain differences in health status according to income (Hertzman, Frank, & Evans, 1994), access was also measured by assessing whether the individual had a regular physician and whether any barriers to care had been encountered in the last 12 months.

Within each income group, frequencies were generated for categorical variables and means were calculated for continuous variables. Chi-square tests were conducted for categorical variables and Goodman and Kruskal’s gamma was used as a measure of association across the income levels. One-way analysis of variance (ANOVA) was conducted on continuous variables and eta used as a measure of association. Eta squared can be considered to represent the proportion of total variability in the dependent variable that can be explained by knowledge of the values of the independent variable (Norusis, 1990). Statistical significance was set a priori at p < .05. All analyses were carried out using SPSS.

RESULTS

Analysis of the data provides insight into the interaction of poverty with environmental, lifestyle, and access determinants to negatively affect health (Tables 1 and 2 and Figure 1). Table 1 shows the unadjusted proportions of adolescents with each characteristic by income category. There was a significant trend toward a lower proportion of adolescents aged 12-14 years in the lower income groups. Under the heading of physical environment, the proportion of adolescents living in a household with a smoker significantly increased as income level declined. Under the heading of social environment, living with a single parent was more likely in the lower income groups, and a high sense of social support was less likely. Under the heading of individual response, mean self-esteem and mean sense of mastery scores were lower in the lower income groups, while mean distress scores were higher. Adolescents in the lower income groups were more likely to be daily smokers and less likely to be happy and interested in life o r physically active. Under the health and function category, the mean number of days spent in bed due to illness was twice as high in the lowest income group compared to the highest. The proportion reporting their self-perceived health as excellent declined significantly from the highest to lowest income groups. Under the heading of health care access, the proportion of adolescents with a regular medical doctor declined significantly from the highest income group (94.7%) to the lowest (78.5%). Also of concern is that the proportion who required but did not receive health care or advice in the last 12 months was higher (8.3%) in the lowest income group compared to the highest (2.2%). Under the heading of health care utilization, the lower income groups were slightly more likely to report not seeing any health professionals in the last 12 months, but those who did see health professionals did so at a higher rate than was observed in the higher income groups.

Table 2 shows the adjusted odds ratios for four categories–low income, disability (bed) days in the past 2 weeks, low self-esteem score, and low sense of mastery–for the characteristics listed. Adjusted odds ratios were calculated using the significant (p < .05) variables remaining in the model following forward stepwise regression analysis.

Low income was associated with increased risk of living in a household with a smoker, living with a single parent, not reporting a high level of social support, self-esteem below the median, being a daily smoker, not having a regular doctor, having been a patient overnight in a hospital in the last 12 months, and having seen a social worker or counsellor in the last 12 months. With the exception of low self-esteem (OR = 1.44), disability days in the past 2 weeks was associated only with illness and health care variables. Low self-esteem was associated with low self-mastery, not being happy and interested in life, less physical activity, and not reporting health as excellent. Low sense of mastery was associated with living with a single parent, not reporting high perceived social support, low self-esteem, high distress, not being happy and interested in life, not self-reporting excellent health, and having a higher number of consultations with a social worker or counsellor in the last 12 months. Low income was not significantly associated with low self-esteem, low sense of mastery, or disability (bed) days after adjusting for the other factors using forward stepwise regression.

Tables 1 and 2 highlight the multidimensional challenge to the health and well-being of adolescents living in low-income households. Figure 1 highlights the associations between mean self-esteem scores, physical activity levels, and household income categories. This figure shows that (1) active adolescents, compared to moderate and inactive adolescents, had a higher mean self-esteem score in all but the highest household income category, and (2) active adolescents in the two lowest income categories had mean self-esteem scores that were about as high as or higher than the mean score for inactive adolescents in the highest household income category. If increasing physical activity levels does in fact improve the self-esteem of adolescents, then policies can be designed to increase the proportion of active adolescents in all household income categories. Such policies can also be designed with particular attention to adolescents living in low-income households, given their vulnerability to low self-esteem as a r esult of other risk factors. Improving self-esteem among adolescents is important given its association with many adverse health and social outcomes.

CONCLUSIONS

Despite public awareness of the negative effects of poverty, the impact of the socioeconomic gradient on the health of young Canadians has remained largely unexplored. This study provides preliminary evidence for an association between income adequacy and several dimensions of well-being. The most important association may be that between poverty, lower levels of physical activity, and lower self-esteem. In addition, this study corroborates earlier studies of the negative effects of low socioeconomic status on several determinants of health.

Hertzman et al. (1994) have noted that “major shifts in the health status of whole populations over time do not necessarily depend upon the implementation of public health or medical control measures against specific diseases. [Studies] point instead to a profound linkage between health and the social environment, including the levels and distribution of prosperity in a society.” Although economic inequality seems to be the underlying influence, there are other immediate factors, such as access to care, which may be more modifiable or amenable to intervention. Several studies have demonstrated the capacity of early intervention to alter the life courses of economically disadvantaged children (Hertzman, 1990). A shift in child health, however, depends largely on policies which cut across a variety of sectors in order to address the full range of determinants which contribute to deficits in both current and future health.

The results of this study signal a need for more focused research into the nature of the relationship between poverty and health among Canadian youth. Follow-up studies are needed to explore the long-term and/or latent effects of socioeconomic factors on health and its determinants. Likewise, the link between material and social deprivation requires a more detailed characterization, and the biological pathway from poverty to poor health needs to be established. Once “strategic times of vulnerability in the life cycle, where … fundamental determinants of health status embed themselves in human biology” (Hertzman, 1990) are identified, the role of public policy in promoting the health of Canadian children will become more clear. Based on the results of this research, focusing an in-depth investigation on the relationship between physical activity and self-esteem may help reveal whether public policies which increase physical activity levels in adolescents can be used to buffer the impact of poverty.

[FIGURE 1 OMITTED]

Table 1

Social and Physical Environment by Income Adequacy for Candian

Adolescents Aged 12 to 19

Income Adequacy (Low [right

arrow] High)

Indicator I II

% age 12-14 years (n=1759) 22.9 (33) 30.3 (80)

Physical Environment

% regular smoker inside 59.0 (85) 58.7 (155)

house (n=1759)

Social Environment

% living with a single parent 22.9 (33) 36.0 (94)

(n=1744)

% living in a rural area 31.1 (32) 34.4 (65)

(n=1182)

% high perceived social 87.7 (114) 81.3 (191)

support (n=1593)

Individual Response

Mean self-esteem scale 19.0 (130) 18.6 (234)

score (n=1594)

Mean mastery scale score 19.0 (123) 18.4 (211)

(n=1443)

Mean distress scale score 4.6 (130) 4.8 (235)

(n=1594)

% “happy and interested in 66.7 (96) 65.4 (172)

life” (n=1756)

% physically active (n=1599) 38.5 (50) 37.3 (88)

% daily smokers (n=1759) 30.6 (44) 24.2 (64)

% drank alcohol in the last 63.2 (91) 56.8 (150)

12 months (n=1759)

Health and Function

% activitiy-limiting injury in 29.2 (42) 27.7 (73)

the last 12 months (n=1759)

Mean number of days spent 1.1 (144) 0.8 (264)

in bed due to illness or injury

in the last 2 weeks (n=1759)

% rating self-perceived 18.8 (27) 28.4 (75)

health as excellent (n=1759)

Disease

% chronic disease diagnosed 48.6 (70) 41.7 (110)

by a health professional

(n=1758)

Mean depression scale score 0.7 (130) 0.8 (234)

(n=1595)

Health Care Access

% with a regular medical 78.5 (113) 87.9 (232)

doctor (n=1759)

% health care or advice 8.3 (12) 5.3 (14)

required but not received in

last 12 months (n=1759)

Health Care Utilization

% consulted health 93.1 (134) 90.9 (240)

professionals in last 12

months (n=1758)

Mean number of consultations 5.1 (144) 3.7 (264)

with a medical doctor in last

12 months (n=1758)

% who have been a patient 12.5 (18) 9.5 (25)

overnight in hospital in the

last 12 months (n=1759)

Mean number of consultations 0.9 (144) 0.9 (264)

with a social worker or cunsellor

in the last 12 months (n=1758)

Mean number of consultations with 0.2 (143) 0.4 (264)

a psychologist in the last 12

months (n=1756)

Income Adequacy (Low [right arrow]

High)

Indicator III IV V

% age 12-14 years (n=1759) 38.0 (205) 34.6 (203) 37.3 (84)

Physical Environment

% regular smoker inside 46.1 (249) 40.6 (238) 27.6 (62)

house (n=1759)

Social Environment

% living with a single parent 20.8(111) 14.2 (83) 8.5 (19)

(n=1744)

% living in a rural area 33.9 (131) 24.7 (94) 26.2 (32)

(n=1182)

% high perceived social 89.9 (436) 90.9 (490) 91.2 (186)

support (n=1593)

Individual Response

Mean self-esteem scale 19.2 (487) 19.6 (539) 19.6 (204)

score (n=1594)

Mean mastery scale score 19.1 (435) 19.5 (496) 20.2 (178)

(n=1443)

Mean distress scale score 4.4 (487) 4.2 (538) 4.4 (204)

(n=1594)

% “happy and interested in 70.9 (383) 76.0 (444) 71.1 (160)

life” (n=1756)

% physically active (n=1599) 36.3 (177) 42.1 (228) 43.1 (88)

% daily smokers (n=1759) 15.7 (85) 12.5 (73) 8.0 (18)

% drank alcohol in the last 51.9 (280) 60.8 (356) 57.3 (129)

12 months (n=1759)

Health and Function

% activitiy-limiting injury in 27.4 (148) 30.5 (179) 32.0 (72)

the last 12 months (n=1759)

Mean number of days spent 0.6 (540) 0.6 (586) 0.5 (225)

in bed due to illness or injury

in the last 2 weeks (n=1759)

% rating self-perceived 28.3 (153) 29.5 (173) 31.6 (71)

health as excellent (n=1759)

Disease

% chronic disease diagnosed 38.3 (207) 44.1 (258) 40.0 (90)

by a health professional

(n=1758)

Mean depression scale score 0.6 (486) 0.5 (541) 0.7 (204)

(n=1595)

Health Care Access

% with a regular medical 86.9 (469) 88.6 (519) 94.7 (213)

doctor (n=1759)

% health care or advice 2.6 (14) 3.4 (20) 2.2 (5)

required but not received in

last 12 months (n=1759)

Health Care Utilization

% consulted health 91.3 (493) 95.4 (558) 95.6 (215)

professionals in last 12

months (n=1758)

Mean number of consultations 3.0 (539) 2.9 (586) 3.0 (225)

with a medical doctor in last

12 months (n=1758)

% who have been a patient 6.9 (37) 3.9 (23) 6.2 (14)

overnight in hospital in the

last 12 months (n=1759)

Mean number of consultations 0.3 (540) 0.2 (586) 0.3 (224)

with a social worker or cunsellor

in the last 12 months (n=1758)

Mean number of consultations with 0.1 (540) 0.1 (584) 0.2 (225)

a psychologist in the last 12

months (n=1756)

Chl-Square/ Measures of

ANOVA Association/

Indicator p-Value p-Value

% age 12-14 years (n=1759) .0064 -.0827/.0241

Physical Environment

% regular smoker inside <.0001 -.2725/<.0001

house (n=1759)

Social Environment

% living with a single parent <.0001 <.3319/<.0001

(n=1744)

% living in a rural area .0326 -.1233/.0082

(n=1182)

% high perceived social .0013 -.1764/.0033

support (n=1593)

Individual Response

Mean self-esteem scale .0001 .1038/<.0001

score (n=1594)

Mean mastery scale score <.0001 .1173/<0001

(n=1443)

Mean distress scale score .1865 -.0549/.0285

(n=1594)

% “happy and interested in .0145 -.1041/.0084

life” (n=1756)

% physically active (n=1599) .2679 -.0721/.0583

% daily smokers (n=1759) <.0001 -.3175/<.0001

% drank alcohol in the last .0219 -.0254/.4760

12 months (n=1759)

Health and Function

% activitiy-limiting injury in .6427 .0506/.1959

the last 12 months (n=1759)

Mean number of days spent .0394 -.0607/.0109

in bed due to illness or injury

in the last 2 weeks (n=1759)

% rating self-perceived .0892 -.0813/.0366

health as excellent (n=1759)

Disease

% chronic disease diagnosed .1366 -.0077/.8304

by a health professional

(n=1758)

Mean depression scale score .3189 -.0309/.2173

(n=1595)

Health Care Access

% with a regular medical .0002 -.2037/.0001

doctor (n=1759)

% health care or advice .0074 -.2322/.0201

required but not received in

last 12 months (n=1759)

Health Care Utilization

% consulted health .0191 -.1995/.0034

professionals in last 12

months (n=1758)

Mean number of consultations <.0001 -.0987/<.0001

with a medical doctor in last

12 months (n=1758)

% who have been a patient .0010 -.2414/.0012

overnight in hospital in the

last 12 months (n=1759)

Mean number of consultations <.0001 -.1178/<.0001

with a social worker or cunsellor

in the last 12 months (n=1758)

Mean number of consultations with .0009 -.0530/.0263

a psychologist in the last 12

months (n=1756)

Table 2

Adjusted Odds Ratios from Forward Stepwise Regression Analysis for

Canadians Aged 12 to 19

Disability (Bed)

Low Income Days in Past 2

Indicator (n=1419) Weeks (n=1419)

Age group: 15-19 compared to 12-14 1.42 (1.10-1.75) –

years

Low income: low (groups I & II) N/A –

compared to middle or higher

(groups III, IV, V)

Physical Environment

Anyone in household smokes 1.76 (1.48-2.03) –

regularly inside the house

Social Environment

Living with a single parent 1.94 (1.63-2.24) –

Living in a rural area N/A N/A

Not reporting high perceived 1.82 (1.45-2.20) –

social support

Individual Response

Self-esteem scale score below 1.32 (1.06-1.58) 1.44 (1.14-1.75)

median

Mastery scale score below median – –

Distress scale score above median – –

Did not report being “happy and – –

interested in life”

Moderate/inactive vs. physically – –

active

Daily smoker 1.47 (1.14-1.80) –

Drank alcohol in the last 12 months – –

Health and Function

Activity-limiting injury in the – 1.53 (1.12-1.84)

last 12 months

Any days spent in bed due to – N/A

illness or injury in the last 2

weeks

Did not rate self-perceived health – –

as excellent

Disease

Chronic disease diagnosed by a – 2.08 (1.76-2.40)

health professional

Depression scale score above median – –

Health Care Access

Does not have regular medical 1.52 (1.15-1.88) –

doctor

Required health care or advice but – 2.31 (1.73-2.89)

not received in the last 12 months

Health Care Utilization

Consulted with any health N/A N/A

professionals in the last 12

months

Number of consultations with a – 1.88 (1.57-2.20)

medical doctor in the last 12

months greater than the median

Was a patient overnight in hospital 1.65 (1.18-2.11) –

in the last 12 months

Number of consultations with a 2.13 (1.70-2.55) –

social worker or counsellor in the

last 12 months greater than the

median

Number of consultations with a – –

psychologist in the last 12 months

greater than the median

Low Self-Esteem Low Sense of

Indicator Score (n=1419) Mastery (n=1419)

Age group: 15-19 compared to 12-14 – 0.72 (0.45-0.98)

years

Low income: low (groups I & II) – –

compared to middle or higher

(groups III, IV, V)

Physical Environment

Anyone in household smokes – –

regularly inside the house

Social Environment

Living with a single parent – 1.50 (1.21-1.80)

Living in a rural area N/A N/A

Not reporting high perceived – 1.64 (1.25-2.02)

social support

Individual Response

Self-esteem scale score below N/A 2.58 (2.35-2.81)

median

Mastery scale score below median 2.54 (2.32-2.77) N/A

Distress scale score above median – 2.12 (1.88-2.36)

Did not report being “happy and 1.87 (1.62-2.12) 2.13 (1.86-2.39)

interested in life”

Moderate/inactive vs. physically 1.33 (1.11-1.56) –

active

Daily smoker – –

Drank alcohol in the last 12 months – –

Health and Function

Activity-limiting injury in the – –

last 12 months

Any days spent in bed due to – –

illness or injury in the last 2

weeks

Did not rate self-perceived health 1.74 (1.48-2.01) 1.32 (1.06-1.58)

as excellent

Disease

Chronic disease diagnosed by a – –

health professional

Depression scale score above median – –

Health Care Access

Does not have regular medical – –

doctor

Required health care or advice but – –

not received in the last 12 months

Health Care Utilization

Consulted with any health N/A N/A

professionals in the last 12

months

Number of consultations with a – –

medical doctor in the last 12

months greater than the median

Was a patient overnight in hospital – –

in the last 12 months

Number of consultations with a – 1.76 (1.30-2.21)

social worker or counsellor in the

last 12 months greater than the

median

Number of consultations with a – –

psychologist in the last 12 months

greater than the median

Note. 95% confidence intervals in parentheses.

REFERENCES

Centre for International Statistics (CIS) on Economic and Social Welfare at the Canadian Council on Social Development. (1994). Countdown 94: Campaign 2000 child poverty indicator report. Ottawa.

Centre for International Statistics (CIS) at the Canadian Council on Social Development. (1995). Child poverty rates, Canada, 1980-1995. http: II www.achilles.net/~council/copvhist.htm.

Evans, R. G., & Stoddart, G. L. (1994). Producing health, consuming health care. In R. G. Evans, M. L. Barer, & T. R. Marmor (Eds.), Why are some people healthy and others not? The determinants of health of populations (pp. 29-64). New York: Aldine DeGruyter.

Frank, J. W., & Mustard, J. F. (1994). The determinants of health from a historical perspective. Daedalus: Health and Wealth, 123(4). http://children.metrotor.on.ca/task/force/must.htm.

Hertzman, C. (1990). Where are the differences that make a difference? (CIAR Population Health Working Paper No. 8). Toronto: Canadian Institute for Advanced Research.

Hertzman, C., & Wiens, M. (1996). Child development and long-term outcomes: A population health perspective and summary of successful interventions. Social Science and Medicine, 43(7), 1083-1095.

Hertzman, C., Frank, J., & Evans, R. G. (1994). Heterogeneities in health status and the determinants of population health. In R. G. Evans, M. L. Barer, & T. R. Marmor (Eds.), Why are some people healthy and others not? The determinants of health of populations (pp. 67-92). New York: Aldine DeGruyter.

Norusis, M. J. (1990). SPSS base system user’s guide. Chicago: SPSS Inc.

Statistics Canada, Health Statistics Division. (1995). National population health survey, 1994-1995 public use microdata files. Ottawa.

Greg Webster and Marian Vermeulen, Central West Health Planning Information Network, Ontario, Canada.

Reprint requests to Thomas J. Abernathy, Central West Health Planning Information Network, 10 George Street, #301B, Hamilton, Ontario, Canada L8P 1C8. Electronic mail may be sent to toma@cwhpin.mcmaster.ca.

COPYRIGHT 2002 Libra Publishers, Inc.

COPYRIGHT 2002 Gale Group