How Do Factors at Different Stages of the Lifecourse Contribute to Oral-health-related Quality of Life in Middle Age for Men and Women?

Abstract

The relative contributions of factors operating in fetal life, childhood, and adulthood to the risk of disease in middle age have become important research issues, but self-perceived oral health has rarely been considered in this context. This study investigated the impact of risk factors operating throughout life on self-perceived oral health, according to the Oral Health Impact Profile (OHIP), at age 50 yrs in 305 individuals from the Newcastle Thousand Families cohort. Factors from early and adult life contributed to the OHIP scores, but in men, self-perceived oral health was mostly explained by factors operating early in life. In women, the number of teeth retained in adulthood had a more prominent impact. Lifecourse influences on oral-health-related quality of life appear different for men and women, which may have implications for the effectiveness of public health interventions and health promotion.

INTRODUCTION

Epidemiological studies have indicated that factors including age, gender, tooth retention, clinical condition, dental attendance, socio-economic status, cultural background, dental anxiety, and smoking may influence oral-health-related quality of life (OHRQoL) (Kelly et al., 2000; Slade, 2002; McGrath and Bedi, 2004; Steele et al., 2004). Indicators of OHRQoL are of growing importance in primary dental care when a patient’s ’burden’ of oral conditions is considered (McGrath and Bedi, 1999a). Such indicators are also useful in the identification of target groups, advocacy for oral health promotion or treatment programs, and service evaluation (Slade and Spencer, 1994a).

A range of validated instruments assessing the impact of oral health on quality of life has been developed, including the Oral Health Impact Profile (OHIP) (Slade and Spencer, 1994b). The 49 questions in the OHIP capture seven conceptually formulated dimensions (functional limitation, physical pain, psychological discomfort, physical, psychological and social disability, and handicap) based on Locker’s (1988) theoretical model of oral health.

Previous studies with the OHIP used for the investigation of influences on OHRQoL have been mostly cross-sectional. The Thousand Families cohort—originally comprised of all 1142 children born in May and June, 1947, to mothers resident within the city of Newcastle upon Tyne in northern England—provides an opportunity to apply a lifecourse approach to OHRQoL, the hypothesis being that different stages of life contribute in different ways to OHRQoL for men and women.

MATERIALS & METHODS

Participants were members of the original cohort who contacted the study team in response to media publicity or were traced through the United Kingdom National Health Service Central Register. Between October, 1996, and December, 1998, self-completion questionnaires on adult health and lifestyle were sent out, and study members were invited to attend for clinical examination. Ethical approval was obtained from appropriate Local Research Ethics Committees. All participants gave their written informed consent.

Early Life Data

The original Newcastle Thousand Families study is described elsewhere (Spence et al., 1954; Miller et al., 1960, 1974). Information on numerous factors—including infant feeding, socio-economic status, birth weight, and housing conditions—was recorded prospectively for all study members and abstracted from existing records. Birth weights were adjusted for gestational age and gender (Leon et al., 1998). Socio-economic status during childhood (I, the most advantaged, II, III, IV, V, the least advantaged) was derived from the UK Registrar General’s Standard Occupational Classification (OPCS, 1990), based on paternal occupation at birth and the occupation of the main wage-earner in the household at age 5 yrs. Housing conditions at birth were scored for the presence of overcrowding, lack of hot water, shared toilet, and dampness or poor repair.

Adult Life Data

Data on adult health and lifestyle were collected by self-completion questionnaire at age 49–51 yrs (Lamont et al., 1998, 2000). Occupational details of the main wage-earner in the household at age 50 were coded to derive social class (OPCS, 1990). The number of pack-years of cigarettes smoked (one pack-year equals one pack of cigarettes smoked per day for one yr) was estimated from the study members’ smoking habits at ages 15, 25, 35, and 50 yrs, as recorded on the self-completion questionnaire at age 50. Cross-sectional data, also recorded on the same questionnaire, on total dietary sugars consumed daily were estimated from the European Prospective Investigation of Causes and Nutrition (EPIC) food frequency questionnaire (Bingham et al., 1997). Similarly, data from the questionnaire were used to derive four categories of weekly alcohol consumption at age 50: none, light (up to 5 units of alcohol for women, 10 for men), moderate (up to 21 units for women and 28 for men) and heavy (Power et al., 1998). The presence of each tooth was recorded by one of two trained research nurses during the clinical examination (Pearce et al., 2004), and by an OHIP questionnaire self-completed at the same time.

The 49 questions in the OHIP asked how frequently individuals had experienced a specific impact during the preceding year. Responses were made on a Likert-type scale (4 = ’very often’, 3 = ’fairly often’, 2 = ’occasionally’, 1 = ’hardly ever’, and 0 = ’never’). The responses were summarized into three indicator variables: the total score (summed) and two frequency scores reporting the number of impacts occurring either ’fairly or very often’ (high frequency) or ’occasionally or more’ (lower and high frequency). The three different indicators describe slightly different things. Total OHIP score represents the overall burden of oral problems; this may be many low-scoring impacts, a few high-scoring impacts, or a mixture. The number of impacts occurring ’fairly or very often’ represents the presence and number of more severe problems. The use of both analyzed in parallel allows for greater insight into the relationships observed.

Statistical Analysis

Explanatory variables were grouped within a conceptual framework (Victora et al., 1997) according to the lifecourse stage at which they would be expected to influence future OHRQoL (Fig.). Eight twins were excluded from all analyses. Gender-specific analyses were carried out, since lifecourse effects on adult health have been suggested to differ by gender for this cohort (Lamont et al., 2000; Parker et al., 2003; Pearce et al., 2004). Gender differences in continuous variables were assessed by the Wilcoxon rank-sum test. Social class, housing conditions, and alcohol consumption were defined as ordinal variables and assessed for trend. Other explanatory variables were treated as continuous variables. The internal reliability of the OHIP items was assessed by Cronbach’s alpha. Relationships between OHIP summary variables and explanatory variables were estimated by multivariable linear regression. Tests for heteroskedasticity and omitted variables were used as additional formal model checks for inspection of residuals and fitted values and the validation of statistical assumptions.

Standardized regression coefficients (b), denoting the increase in OHIP indicators for a standard deviation increase in an explanatory variable, are presented with 95% confidence intervals (CI).

The percentage of total variance in each OHIP indicator directly accounted for by each lifecourse stage was estimated by the difference in R² (%) between regression models with and without the group of variables in the lifecourse stage of interest (Fig.), and including all other variables in the regression model. The overall effect of early life, including indirect effects mediated through factors operating in adulthood, was estimated by the value of R² from a regression model containing those variables operating in early life alone. The overall contribution of adult variables, including effects mediated through the number of retained teeth, was estimated by the difference in R² (expressed as a percentage) between models with and without this group of variables, and including all other variables except number of retained teeth. Tooth retention was considered separately, since factors affecting tooth loss operate throughout life, although the most significant period appears to be adulthood (Pearce et al., 2004). Study members with missing data were included in all analyses for which they contributed complete data, except for the hierarchical modeling, where study members with any missing data were excluded to allow for a comparison between different regression models. The statistical software package Stata was used for all analyses (STATA, Release 8.0, StataCorp, College Station, TX, USA).

RESULTS

Of the original 1142 study members recruited in 1947, 832 (72%) were traced at age 50. Of these, 574 (69%) completed the health and lifestyle questionnaire, and 305 (27%) (127 men, 178 women) of the 329 study members attending the dental examination fully completed the OHIP questionnaire. This sample did not differ significantly from the members of the original cohort not included in this analysis, in terms of any of the factors acting at birth [standardized birth weight (p = 0.4), socio-economic status (p = 0.4), and housing (p = 0.3)] and during infancy and childhood (socioeconomic status, p = 0.8), with the exception of gender (Pearce et al., 2004). There were more women in this sample than men compared with the original cohort (p > 0.05). Descriptive statistics are given in Tables 1 and 2 . Women reported a significantly higher median total OHIP score, and a greater number of OHIP impacts occurring ’occasionally or more’, than did men (Table 1 ). Internal consistency between OHIP items was good, with Cronbach alpha values greater than 0.94.

Predictors of Oral-health-related Quality of Life

Men

Social class at birth (b = 3.48, 95%CI 0.36, 6.59, p = 0.031) and number of retained teeth (b = −0.57, 95%CI −0.95, −0.18, p = 0.004) were significantly associated with total OHIP score. However, in the multivariable model, only the number of teeth remained significant (p = 0.01). Social class at birth and at age 5 yrs (b = 0.69, 95%CI 0.24, 1.14, p = 0.003) and housing conditions at birth (b = 0.53, 95%CI 0.07, 0.98, p = 0.024) were associated with the number of OHIP impacts occurring ’often’. The number of retained teeth was the only significant factor predictor of the number of OHIP impacts occurring ’occasionally or more’ (b = −1.69, 95%CI −2.72, −0.67, p = 0.002).

Women

The number of retained teeth showed a significant relationship with total OHIP score (b = −8.65, 95%CI −11.8, −5.5, p < 0.001), the number of OHIP impacts occurring ’often’ (b = −1.24, 95%CI −1.84, −0.63, p < 0.001), and the number occurring ’occasionally or more’ (b = −2.99, 95%CI 0.09, 1.36, p < 0.001), and remained so after adjustment for all other co-variates (p < 0.001). Number of pack-years of cigarettes smoked was significantly associated with the number of OHIP impacts occurring ’occasionally or more’ (b = 1.22, 95%CI 0.09, 2.36, p = 0.037). However, this was not independent of the number of retained teeth (p = 0.55). The number of OHIP impacts occurring ’often’ increased as breast-feeding duration increased (b = 0.73, 95% CI 0.09, 1.36, p = 0.026), and remained significant in the adjusted model (p = 0.008). No other early-life factors predicted any of the OHIP variables.

Relative Contribution of Each Lifecourse Stage

The explanatory variables included in this study explained between 17 and 21% of the total variation in OHIP indicators in men, and 24 and 29% in women (Table 3 ). The greatest proportion of variation in total OHIP score in men was explained by factors acting in early life, including the indirect effects mediated through factors operating later in life. Total OHIP score in women was mostly explained by the direct impact of number of retained teeth, which directly explained only 1% of variation in men.

In both genders, early-life factors accounted for the greatest proportion of variation in the number of OHIP impacts occurring ’fairly or very often’. The number of retained teeth again directly explained a greater amount of variation in women (9%) than in men (0.5%).

For the number of impacts occurring ’occasionally or more’, adult socio-economic position and lifestyle, including the indirect effects mediated through the number of retained teeth, accounted for the greatest proportion of variation in men. The direct impact of tooth retention explained the most variation in women (15%), although, again, it had a negligible effect in men (1%).

DISCUSSION

The total proportion of variation in OHRQoL was generally high, ranging from 17 to 29% in the six models constructed. However, there is inevitably much variation that cannot be accounted for by the variables included in this investigation. It is possible that the unexamined factors that may account for the unexplained variation within these analyses may also vary between the genders.

Longitudinal studies of this nature allow for comparisons of the characteristics of the initial and final samples, and the identification and management of any major biases. The study members participating in this investigation did not differ significantly from the remainder of the original cohort with respect to any factor in early life other than gender, by which separate analyses were undertaken. In addition, inclusion of cohort members who had moved out of the study region increased the representativeness of the population studied. We are therefore confident that the participants examined were representative of the whole cohort, although the possibility of selection bias cannot be completely ruled out.

Women generally reported poorer OHRQoL than men for each OHIP measure, consistent with previous findings on a national sample of a similar age group (Kelly et al., 2000). However, the lifecourse influences on OHRQoL also differed by gender, and varied according to the frequency of impact. In women, only factors acting during adulthood, particularly the number of retained teeth, had a significant impact on total OHIP score and the number of OHIP impacts occurring ’often’. In men, those of a more advantaged social class at birth and with fewer lost teeth in adulthood tended to have the lowest total OHIP scores, reflecting a lower negative impact of oral health on daily activities. More advantaged social class at birth and during childhood and better housing conditions at birth were also related to fewer OHIP impacts occurring ’often’. Tooth retention in adulthood was associated with fewer occasional impacts.

The greatest proportion of variation in total OHIP score was explained by early-life factors in men, but by tooth retention in women. Early life accounted for the most variation in the number of frequently occurring OHIP impacts for both genders, but when impacts occurring ’occasionally or more’ were considered, factors acting during adulthood explained the most variation in men, and the number of teeth accounted for the most variation in women.

The direct impact of tooth retention on each OHIP measurement differed between men and women. Women’s OHRQoL was more greatly affected than that of men by problems that were only occasional. In contrast, early-life factors had the most prominent impact in men.

Although tooth retention is a significant predictor of OHIP scores in men, they appear to be more influenced by their background in terms of the variation in OHIP score. This may be as much a reflection of the kind of care they sought to address dental problems, as of underlying oral health priorities. Women appear to be more influenced directly by the number of teeth they have.

Inglehart et al.’s review of the limited literature (2002) identified several ways in which gender differences may be related to health and OHRQoL. McGrath and Bedi (1999b) specifically identified differences between men and women in the way that quality of life is affected by oral health in older British adults, and proposed that OHRQoL is not gender-free. This study suggests that models of the lifecourse influences on OHRQoL in late adulthood may be substantially different between men and women. It would be easy to interpret this in terms of stereotypes, but different lifecourse influences in OHRQoL between men and women may be of importance. Promotion of a healthier adult lifestyle and continued improvements in dental prevention to increase tooth retention would appear to be the public health interventions most likely to improve self-perceived oral health in women. However, such measures may have a lesser impact on men, who appear to be less directly affected by tooth retention. For both genders, social background may be an important correlate with the impact of oral health on daily living in later life, particularly when frequently occurring impacts are considered. An investigation of the impact of each lifecourse stage, on each of the seven specific OHRQoL domains defined by the OHIP, may help us to identify the mechanisms leading to the apparent gender differences.

Table 1.

Descriptive Statistics for Continuous Variables

	Men (N = 120)		Women (N = 161)
Variable	Range	Mean (SD)	Range	Mean (SD)	P
^a Information on infant feeding was available for 118 men and 156 women.
^b Excluding 41 men and 76 women who had never smoked.
^c Dietary data were available for 118 men and 154 women.
* IQR = Interquartile range.
Birth weight (kg)	1.9, 4.6	3.4 (0.5)	2.1, 4.5	3.4 (0.5)	0.22
	Range	Median (IQR)*	Range	Median (IQR)
Duration of breast feeding (days)^a	0, 411	56 (28–207)	0, 430	62 (21–198)	0.88
Cigarette smoking (pack-years)^b	0.2, 73	23 (8–35)	0.1, 42	19 (7–26)	0.002
Total dietary sugars (g/day)^c	32, 247	117 (86–143)	34, 269	118 (87–143)	0.62
No. of teeth	0, 32	25 (19–27)	0, 32	24 (19, 27)	0.49
Total OHIP Score	1, 97	15 (9–27)	2, 123	22 (12–40)	0.002
No. of OHIP impacts occurring often	0, 19	0 (0–1)	0, 22	0 (0–2)	0.05
No. of OHIP impacts occurring occasionally or more	0, 31	4 (2–8)	0, 38	6 (3–12)	0.001

Table 2.

Descriptive Statistics for Categorical Variables

		Men		Women
Variable		Number	%	Number	%
^a As defined by the UK Registrar General’s Standard Occupational Classification, where I is the most advantaged.
^b None to three or more of: overcrowding, lack of hot water, shared toilet; and dampness or poor repair.
Social class at birth^a	I,II	15	12.5	17	10.6
	III	70	58.3	104	65.0
	IV,V	35 (n = 120)	29.1	39 (n = 160)	24.4
Social class at age 5 yrs^a	I,II	9	7.5	16	10.1
	III	72	60.0	93	58.5
	IV,V	39 (n = 120)	32.5	50 (n = 159)	31.5
Social class in adulthood^a	I,II	54	45.0	84	53.9
	III	47	39.2	44	28.2
	IV,V	19 (n = 120)	15.8	28 (n = 156)	17.9
Housing conditions (score at birth^b)	0	58	48.3	85	52.8
	1	31	25.8	39	24.2
	2	18	15.0	19	11.8
	3 or more	13 (n = 120)	10.9	18 (n = 161)	11.2
Alcohol	None	9	7.5	28	15.0
	Light	52	43.3	69	38.8
	Moderate	46	38.3	69	40.0
	Heavy	13 (n = 120)	10.9	11 (n = 160)	6.3

Table 3.

Direct and Overall Contributions to Explaining the Variation in OHIP Scores in Men and Women at ages 49–51 yrs from Circumstances and Experiences at Each Stage in the Lifecourse

		Percentages of Variation Explained
		Total OHIP Score		No. of OHIP Impacts Occurring Fairly Often or Often		No. of OHIP Impacts Occurring Occasionally or More
Source		Men (n = 116)	Women (n = 146)	Men (n = 116)	Women (n = 146)	Men (n = 116)	Women (n = 146)
^a Ignoring effects mediated through number of teeth.
^b Cigarette smoking, alcohol consumption, total sugars. The overall contribution of early-life variables includes indirect effects mediated through adult socio-economic position and lifestyle.
Fetal life (birth weight, socio-economic position, and housing score at birth)	(Direct)	7.3	3.6	9.7	4.8	5.1	3.1
Infancy and childhood (breast-feeding, socio-economic position)	(Direct)	1.0	3.5	2.6	4.3	3.7	2.4
Total early life (including indirect effects)	(Overall)	8.5 (7.9)^a	6.1 (10.0)^a	18.5 (17.2)^a	9.2 (12.9)^a	6.5 (6.5)^a	5.0 (8.5)^a
Adult socio-economic position and lifestyle^b	(Direct)	4.8	3.6	1.9	2.3	8.8	4.7
Total adult (including indirect effects mediated through number of retained teeth)	(Overall)	6.7	6.7	2.4	6.4	12.1	7.8
Number of retained teeth	(Direct)	1.4	15.9	0.5	8.9	1.4	15.4
Total		16.7	28.7	21.4	24.5	20.1	28.1

Figure.

Conceptual framework for fetal, infant, childhood, and adult influences on oral-health-related quality of life at ages 49–51 yrs. The solid lines represent the direct effects of each stage of the lifecourse on oral-health-related quality of life. The dashed lines represent indirect effects mediated through later stages of the lifecourse.

Footnotes

Notes

Acknowledgements

This work formed part of the BMedSci degree completed by JM and was funded through a bursary from The Health Foundation. We thank all the Thousand Family Study members for taking part in this study, and the study teams past and present. We thank the Wellcome Trust, the Minnie Henderson Trust, the Sir John Knott Trust, and the Special Trustees of the Newcastle Hospitals for funding. Jan Gebbie and Jean Gerrard carried out the oral health examinations. EPIC food frequency questionnaires were analyzed by the MRC Dunn Clinical Nutrition Centre, Cambridge.

References

Bingham SA, Gill C, Welch A, Cassidy A, Runswick SA, Oakes S, et al. -1997-. Validation of dietary assessment methods in the UK arm of EPIC using weighed records, and 24-hour urinary nitrogen and potassium and serum vitamin C and carotenoids as biomarkers. Int J Epidemiol 26-Suppl 1-:S137–S151.

Inglehart MR, Silverton SF, Sinkford JC -2002-. Oral health-related quality of life: does gender matter? In: Oral health-related quality of life. Inglehart MR, Bagramian RA, editors. Chicago: Quintessence Publishing, pp. 111–121.

Kelly M, Steele J, Nuttall N, Bradnock G, Morris J, Nunn J, et al. -2000-. Adult dental health survey: oral health in the United Kingdom 1998. London: The Stationery Office.

Lamont DW, Parker L, Cohen MA, White M, Bennett SM, Unwin NC, et al. -1998-. Early life and later determinants of adult disease: a 50 year follow-up study of the Newcastle Thousand Families cohort. Public Health 112:85–93.

Lamont D, Parker L, White M, Unwin N, Bennett SMA, Cohen M, et al. -2000-. Risk of cardiovascular disease measured by carotid intima-media thickness at age 49–51: lifecourse study. BMJ 320:273–278.

Leon DA, Lithell HO, Vagero D, Koupilova I, Mohsen R, Berglund L, et al. -1998-. Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15000 Swedish men and women born 1915–29. BMJ 317:241–245.

Locker D -1988-. Measuring oral health: a conceptual framework. Community Dent Health 5:3–18.

McGrath C, Bedi R -1999a-. The value and use of ’quality of life’ measures in the primary dental care setting. Prim Dent Care 6:53–57.

McGrath C, Bedi R -1999b-. The importance of oral health to older people’s quality of life. Gerodontology 16:59–63.

10.

McGrath C, Bedi R -2004-. The association between dental anxiety and oral health-related quality of life in Britain. Community Dent Oral Epidemiol 32:67–72.

11.

Miller FJW, Court SDM, Walton WS, Knox EG -1960-. Growing up in Newcastle upon Tyne. London: Oxford University Press.

12.

Miller F, Court S, Knox E, Brandon S -1974-. The school years in Newcastle upon Tyne, 1952–62. London: Oxford University Press.

13.

Office for Population Censuses and Surveys -OPCS- -1990-. Standard occupational classification. Vols. 1,2,3. London: HMSO.

14.

Parker L, Lamont DW, Unwin N, Pearce MS, Bennett SM, Dickinson HO, et al. -2003-. A lifecourse study of risk for hyper-insulinaemia, dyslipidaemia and obesity -the central metabolic syndrome- at age 49–51 years. Diabet Med 20:406–415.

15.

Pearce MS, Steele JG, Mason J, Walls AW, Parker L -2004-. Do circumstances in early life contribute to tooth retention in middle age? J Dent Res 83:562–566.

16.

Power C, Rodgers B, Hope S -1998-. U-shaped relation for alcohol consumption and health in early adulthood and implications for mortality [letter]. Lancet 352:877.

17.

Slade GD -2002-. Assessment of oral health related quality of life. In: Oral health-related quality of life. Inglehart MR, Bagramian RA, editors. Chicago: Quintessence Publishing, pp. 29–45.

18.

Slade GD, Spencer AJ -1994a-. Social impact of oral conditions among older adults. Aust Dent J 39:358–364.

19.

Slade GD, Spencer AJ -1994b-. Development and evaluation of the oral health impact profile. Community Dent Health 11:3–11.

20.

Spence J, Walton WS, Miller FJW, Court SDM -1954-. A thousand families in Newcastle upon Tyne. London: Oxford University Press.

21.

Steele JG, Sanders AE, Slade GD, Allen PF, Lahti S, Nuttall N, et al. -2004-. How do age and tooth loss affect oral health impacts and quality of life? A study comparing 2 national samples. Community Dent Oral Epidemiol 32:107–114.

22.

Victora CG, Huttly SR, Fuchs SC, Olinto MT -1997-. The role of conceptual frameworks in epidemiological analysis: a hierarchical approach. Int J Epidemiol 26:224–227.