Association of objective health factors with self-reported health.

OBJECTIVES: There is a strong relationship between subjective health and mortality, level of functional ability and medical consumption. The aim of this study was to describe the correlation of objective health-related factors with self-reported health (SRH) of a sample of the Belgian population.
METHODS: Participants were recruited during an exhibition at the Brussels Exhibition Centre. They completed a visual analogue scale assessing their SRH. Medical history and health related parameters of the participants were recorded.
RESULTS: In total 974 visitors participated. From the multivariate analysis we found an association between low SRH and diabetes (OR 0.23-0.80), increased body mass index (OR 0.52-0.74), coronary heart disease (OR 0.28-0.97), smoking (OR 0.38-0.89), speaking Dutch (OR 0.40-0.92), not knowing length (OR 0.36-0.99), family history of breast cancer (OR 0.41-0.94), family history of coronary heart disease (OR 0.45-095) and aging (OR 0.84-0.99). Following a cholesterol-lowering diet was associated with a high SRH (OR 1.10-2.44).
CONCLUSIONS: Most of the factors associated with low SRH are known and confirm what has previously been reported in literature. However, the associations between low SRH and not knowing your length, speaking Dutch or having afamily history of breast or colon cancer, as well as the association between high SRH and being on a cholesterol-lowering diet are interesting new findings.

Introduction

Good health seems to be an important determinant for the well-being and quality of life of people. The context of health is not limited to the physical well-being of individuals or communities but should take into account the emotional, socio-economic, mental, spiritual and cultural well-being of the individuals in the community. The most widely used definition of health is that provided by the World Health Organization (WHO): "Health is a complete state of physical, mental and social wellbeing, and not merely the absence of disease or infirmity" [1].

We aimed to measure the subjective health status in one question, which means that our measurements included not only physical health, but also social and psychological health components. We may assume that such a subjective health status will be affected by the presence of symptoms or specific complaints, but also by the medical diagnoses and risk factors known by the participant [2].

In the past, several studies have been set up to identify the determinants of subjective health. However, it has never been possible to develop an accurate conceptual description of subjective health status [3].

According to the second goal of the WHO Millennium Development Goals, everyone should have the possibility to develop one's own health potential. The subjective assessment of one's own health is considered to be a good indicator in this context, both at the individual level and at the level of society.

Many studies were able to demonstrate a strong cohesion between subjective health and mortality [4]. Subjective health also seems to be a good predictor of morbidity, the level of functional ability and medical consumption [5-10]. The subjective perception of health status is therefore a useful tool to detect high-risk persons and to estimate care requirements.

This study aimed to describe the correlation of objective health-related factors with the perceived health of a sample of the Belgian population.

Methods

PARTICIPANTS

Participants were recruited during a food exhibition at the Brussels Exhibition Centre from October 6th to October 21th 2012. All adult visitors were invited to participate. Exclusion criteria were: pregnancy, taking vitamin K antagonists, showing signs of addiction to alcohol, medication or drugs, or being intolerant to blood and / or finger pricks.

QUESTIONNAIRE

To evaluate the self-reported health (SRH) the part of the EQ-5D questionnaire measuring SRH using a visual analogue scale (VAS) was used [11]. We did not use the other scales from the EQ-5D questionnaire. Participants were asked to score their perceived health status on a scale between 0 and 100 with 0 corresponding with the worst health participants could possibly imagine and 100 meaning the best imaginable health.

Scoring on the VAS was performed after recording the participants' age, gender and zip-code but before other health-related questions were asked.

Participants were asked if they knew their length, weight, abdominal circumference, blood pressure, cholesterol level and blood sugar level. Subsequently these parameters were also measured. Furthermore, participants were asked about their medical history (coronary heart disease, hypercholesterolemia, diabetes, hypertension, other diseases) and their family history (breast cancer, coronary heart disease, diabetes, colon cancer). They were also asked about their latest tetanus vaccination.

MEASUREMENTS

Some parameters of the enrolled subjects were measured. We respectively used a digital personal scale Seca Sensa 804 to measure weight, a Seca 206 wall-mounted measuring tape for height, a Seca 201 ergonomic circumference measuring tape for abdominal circumference, a calibrated DS-54 WelchAllyn sphygmomanometer blood pressure device for blood pressure, a OneTouch device using capillary blood for blood sugar and an Accutrend Plus monitor using capillary blood for total cholesterol. Capillary blood was obtained by pricking the index finger. A short medical history was obtained, focusing on food and beverage intake during the two hours before the measurement.

APPROVAL OF THE ETHICAL COMITY

The study protocol was approved by the ethical committee of the University Hospital Brussels. Visitors of the exhibition were allowed to participate after they had read the patient-information leaflet and had signed an informed consent form. After completing the questionnaire, participants were offered personalised health advice.

STATISTICAL PROCESSING

Data were anonymously recorded in a mySQL database using an online custom-made PHP-based interface, hosted by the Faculty of Medicine and Pharmacy of the Vrije Universiteit Brussel. Incomplete data sets were eliminated from the database. From the entered data, body mass index (BMI) was generated by the system. Cardiovascular risk was estimated by the system using the Belgian SCORE risk tables [12]. Participants were assigned to either a low-SRH-group or a high SRH-group, using the SRH median (7.14) as an arbitrary cutoff. This permitted us to compare two groups of a similar size.

Statistical analyses were performed using SPSS 22, using cross-tables and the Chi-Square test for discrete variables. A t-test was used to compare the means of two groups and the one-way ANOVA was used to compare the means of three groups.

A logistic regression was performed to determine variables linked with high or low SRH. The following variables were entered: gender, age-groups, language, region, hypertension, hypercholesterolemia, diabetes, coronary heart disease, no disease, family history of coronary heart disease, family history of colon cancer, family history of diabetes, family history of breast cancer, no family history, tetanus vaccination up-to-date, weight, length, abdominal circumference, does know blood pressure, does know blood sugar level, does know cholesterol level, no treatment or diet for cholesterol, diet for cholesterol, statin for cholesterol, plant stanol for cholesterol, smoker, alcohol abuse, physical activity, BMI (4 groups), cardiovascular risk groups (SCORE low, intermediate and high).

Results

STUDY POPULATION

In total 974 visitors participated: 31% men and 69% women. Ages ranged from 18 to 90 years, with an average of 53.3 years. Most of the participants (77%) were from the Flemish region, 18% from the Brussels Region and 5% from the Walloon Region.

The median SRH was 71.4%: 499 participants had a SRH of 71% or lower and 475 had a SRH of more than 71%.

The mean SRH was the highest among the very young (< 20 year) and decreased gradually until the age group of 50 to 59 years, to remain stable in the older age groups (Fig. 1).

Fig. 1.

Mean self-reported health (SRH) per age-group.

MEAN SELF-REPORTED HEALTH

The mean SRH did not differ between men (72%) and women (71%) (Tab. I.). Neither was there a significant difference between regions, although Dutch-speaking participants had a lower SRH than French-speaking participants.

Tab. I.

Mean self-reported health (SRH) per group.

	N	Mean SRH	Std. Dev.	p-value
Men Women	299 675	72.36 71.00	15.113 13.314	0.160
Brussels Region Flemish region Walloon region	174 749 51	71.36 71.19 74.94	17.274 13.030 13.143	0.176*
Dutch French	857 117	70.75 76.34	13.746 14.069	< 0.001
No hypertension Hypertension	749 225	72.44 68.04	13.148 15.707	< 0.001
No hypercholesterolemia Hypercholesterolemia	656 318	71.98 70.27	13.891 13.860	0.072
No diabetes Diabetes	912 62	71.85 65.11	13.592 16.686	< 0.001
No coronary heart disease Coronary heart disease	919 55	72.02 61.35	13.236 19.798	< 0.001
No disease Some disease	528 446	73.15 69.38	12.941 14.704	< 0.001
No family history of coronary heart disease Family history of coronary heart disease	824 150	72.11 67.65	13.466 15.578	< 0.001
No Family history of colon cancer Family history of colon cancer	874 100	71.55 70.33	14.015 12.840	0.376
No family history of diabetes Family history of diabetes	736 238	72.13 69.23	13.417 15.107	0.005
No family history of breast cancer Family history of breast cancer	861 113	71.67 69.55	14.000 12.990	0.109
No Family history of no disease Family history of some disease	527 447	72.83 69.76	13.393 14.308	0.001
Tetanus vaccination not up-to-date or unknown Tetanus vaccination up-to-date or unknown	422 552	71.32 71.49	13.308 14.343	0.849
Does not know weight Does know weight	21 953	69.57 71.46	12.424 13.931	0.499
Does not know length Does know length	78 896	67.68 71.75	16.289 13.631	0.035
Does not know abdominal circumference Does know abdominal circumference	892 82	71.62 69.23	13.666 16.135	0.197
Does not know blood pressure Does know blood pressure	140 834	73.31 71.10	12.662 14.076	0.062
Does not know blood sugar level Does know blood sugar level	321 653	72.29 70.99	14.058 13.808	0.174
Does not know cholesterol level Does know cholesterol level	330 644	72.36 70.94	14.311 13.666	0.136
No diet or treatment for cholesterol Diet or treatment for cholesterol	622 352	71.72 70.89	13.926 13.849	0.369
No diet for cholesterol	846	71.28	13.996	0.390
Diet for cholesterol	128	72.37	13.241
No statin treatment for cholesterol Statin treatment for cholesterol	794 180	72.12 68.32	13.553 14.974	0.001
No plant stanol treatment for cholesterol Plant stanol treatment for cholesterol	822 152	71.14 72.93	14.111 12.617	0.117
Non-smokers Smokers	865 109	71.75 68.77	13.897 13.676	0.034
Non-alcohol abusers Alcohol abusers	944 30	71.55 67.47	13.784 16.870	0.200
No physical activity Physical activity	678 296	70.78 72.90	14.353 12.692	0.029
Underweight (BMI < 18.5) Normal weight (18.5 < BMI < 25) Overweight (25 < BMI < 30) Obesity (BMI > 30	39 433 360 142	74.74 73.47 70.87 65.66	10.371 14.231 13.167 13.846	< 0.001*
Low cardiovascular risk SCORE Intermediate cardiovascular risk SCORE High cardiovascular risk SCORE	690 149 135	72.67 71.88 64.51	13.095 12.054 17.430	< 0.001*

p = one way Anova

SRH was lower among participants with known hypertension, diabetes and coronary heart disease compared to those without these conditions.

Having a disease was related to a lower SRH than not having a disease. Having a family history of diabetes or coronary heart disease was also related to a lower SRH. However, having a family history of breast cancer or colon cancer was not related to a lower SRH. But having a family history of a hereditary disease was related to a lower SRH than not having family history of such a disease.

Several other health parameters were assessed but only few were related to a lower SRH: not knowing one's own length, taking a statin, smoking and having little physical activity.

There was a linear relationship between BMI and SRH: the higher the BMI the lower the SRH. A similar relationship was found between cardiovascular risk and SRH: the higher the risk, the lower the SRH.

MEAN VALUES FOR PARAMETERS

In the group with low SRH, median age, mean previous and actual systolic blood pressures, previous and actual weights, previous and actual blood sugar levels, actual BMI and actual abdominal circumferences were higher than in the high SRH group. In the group with high SRH, mean previous and actual lengths were higher (Tab. II).

Tab. II.

Mean values for parameters for low and high self-reported health (SRH).

	SRH group	N	Mean	Std. Dev.	p-value
Age	Low High	499 475	55.39 51.07	16.600 18.156	< 0.001
Last measured systolic blood pressure	Low High	355 304	127.47 124.69	16.488 11.734	0.014
Last measured diastolic blood pressure	Low High	355 304	78.26 77.86	9.245 7.735	0.54
Last measured weight	Low High	483 466	73.62 69.63	15.651 12.465	< 0.001
Last measured length	Low High	442 448	166.52 168.06	8.741 8.594	0.008
Last measured abdominal circumference	Low High	30 19	95.63 90.21	12.164 13.559	0.165
Last measured blood sugar	Low High	261 258	92.17 85.91	17.741 10.267	< 0.001
Last measured cholesterol	Low High	265 254	196.26 191.00	34.851 35.787	0.091
Mean number of cigarets per day	Low High	499 475	1.62 1.19	5.150 4.614	0.169
Mean number of alcoholic beverages per day	Low High	499 475	4.29 3.81	7.673 5.866	0.277
Actual systolic blood pressure	Low High	499 475	125.90 124.00	16.175 13.773	0.049
Actual diastolic blood pressure	Low High	499 475	77.58 77.68	9.240 7.933	0.861
Actual blood sugar	Low High	499 475	109.70 98.71	53.936 22.465	< 0.001
Actual body weight	Low High	499 475	73.56 69.62	15.633 12.412	< 0.001
Actual length	Low High	499 475	166.12 167.85	8.792 8.789	0.002
Actual total cholesterol	Low High	499 475	181.02 179.43	33.648 33.501	0.460
Actual abdominal circumference	Low High	499 475	92.82 87.37	13.878 12.410	< 0.001
Actual body mass index	Low High	499 475	26.062 24.226	4.8485 3.9512	< 0.001

LOGISTIC REGRESSION

Where possible, parameters were dichotomized and included in a logistic regression (Tab. III.). Participants were divided into age groups of 10 years, four BMI groups and three cardiovascular risk groups. A multivariate analysis confirmed the relationships between low SRH and diabetes (OR 0.23-0.80), increased BMI (OR 0.52-0.74), coronary heart disease (OR 0.28-0.97), smoking (OR 0.38-0.89), speaking Dutch (OR 0.40- 0.92), not knowing length (OR 0.36-0.99), family history of breast cancer (OR 0.41-0.94), family history of coronary heart disease (OR 0.45-095) and aging (OR 0.84-0.99). Following a cholesterol-lowering diet was associated with a high SRH (OR 1.10-2.44).

Tab. III.

Logistic regression: factors related to high or low self-reported health (SRH).

	B	p-value	OR	95% C.I.for OR
				Lower	Upper
Age-groups (per 10 years)	-0.092	0.037	0.912	0.836	0.995
Speaking Dutch (vs French)	-0.501	0.019	0.606	0.399	0.920
Diabetes (Y/N)	-0.839	0.008	0.432	0.232	0.803
Coronary heart disease (Y/N)	-0.658	0.041	0.518	0.275	0.974
Family history of coronary heart disease (Y/N)	-0.421	0.027	0.656	0.451	0.954
Family history of breast cancer (Y/N)	-0.479	0.025	0.619	0.407	0.942
Does not know length (Y/N)	-0.521	0.047	0.594	0.355	0.992
Does not know blood sugar level (Y/N)	-0.503	0.058	0.605	0.360	1.017
Does not know cholesterol level (Y/N)	0.483	0.077	1.621	0.949	2.770
Follows a cholesterol diet (Y/N)	0.491	0.016	1.634	1.097	2.435
Smoking (Y/N)	-0.547	0.013	0.579	0.377	0.889
Body mass index (4 groups underweigh > obesity)	-0.476	<0.001	0.621	0.520	0.742

Variable(s) entered on step 1: gender, age-groups, language, region, hypertension, hypercholesterolemia, diabetes, coronary heart disease, no disease, family history of coronary heart disease, family history of colon cancer, family history of diabetes, family history of breast cancer, no family history, tetanus vaccination up-to-date, weight, length, abdominal circumference, does know blood pressure, does know blood sugar level, does know cholesterol level, no treatment or diet for cholesterol, diet for cholesterol, statin for cholesterol, plant stanol for cholesterol, smoker, alcohol abuse, physical activity, body mass index (4 groups), cardiovascular risk groups (SCORE low, intermediate and high).

Discussion

SAMPLE POPULATION

Women and participants from the Flemish region were overrepresented in our study. Extra bias was caused by the fact that all participants were visitors of a food exhibition, meaning that severely disabled or seriously ill people were less likely to participate. As we never aimed to include a representative sample of the Belgian population, this did not hamper the interpretation of our results. Our purpose was to describe the correlation of objective health-related factors with SRH in an arbitrary sample of the Belgian population.

SELF-REPORTED HEALTH

Health status was not assessed by an objective third party. Even though self-assessment is undoubtedly influenced by external factors, including the views of other people, it was ultimately the participant himself who answered the questions.

This type of subjective assessment could rather be an emotional reflection than a systematic, cognitive analysis. Moreover, the subjective measurement of health is without any doubt related to the participants' quality of life.

We were particularly interested in "general" health and not in "current" health. With this subtle difference we tried to reduce the influence of temporary health issues. However, it is not clear how well participants were able to distinguish their general health from their actual health.

The median SRH in our study was 71.4%. In a Finnish study, a similar population had a SRH of 70% [13]. A similar study in Singapore also reported a SRH of approximately 70% [14]. Another study among institutionalised elderly reported a SRH of 78 [15]. Some caution is needed while comparing SRH-results from different studies and different countries, because a subjective approach of health could be highly influenced by cultural diversity. However, such a subjective assessment is sometimes influenced by some cultural related tendencies to complain more or to reflex a rather pessimistic view. Also the functional status appears to be related to ethnic variation [3]. Thereupon, SRH seems to be more useful to monitor health over a period of time or before and after a treatment.

Another limitation of the SRH assessment using a VAS is the difficulty to differentiate participants considering themselves in good or bad health because every participant might use its own arbitrary cut-off between good and bad health. Assessment methods as proposed by the WHO or the one used in the National Health Interview Survey in de United States do not have this disadvantage [10, 16].

COMPARISON WITH OTHER STUDIES

In the Belgian Health Interview Survey (HIS) of 2008 – using the assessment method proposed by the WHO – 23% of the participants considered themselves not to be in good health [17]. As we used a VAS we cannot compare this figure with our results. However, some comparisons are possible. In the HIS more women (25%) than men (20%) reported a bad health. This was not confirmed by our findings. Neither could we confirm that inhabitants of the Flemish region complained less about bad health (21%) than the inhabitants of the Walloon region (26%) or Brussels region (26%).

In our study (very) young participants scored the highest mean SRH with a gradual decrease until the age of 50. The HIS showed a similar evolution. But contrary to our findings that SRH remained stable in the older age groups, the HIS reported a continuing decrease in SRH after the age of 60. This can be explained by the fact our study population was a selected "mobile" population and that disabled or seriously ill people were less likely to visit the exhibition.

Another reliable source on health is the Organisation for Economic Co-operation and Development (OECD) [18]. From the most recent data (2011), 74% of the Belgian adult population considers its health as good. In Portugal, for example, only 50% of the population considers its health as good. The highest SRH is found in the United States with 90%. The figures for Belgium are comparable with The Netherlands (76%) although some other West-European countries such as Germany (65%) and France (68%) report substantially lower figures.

FACTORS ASSOCIATED WITH SELF-REPORTED HEALTH

This multivariate analysis reconfirmed the relationship between low SRH and known diabetes, as has already been reported in other studies [19, 20]. Several studies also found increased BMI to be related to low SRH [21-23]. Similar evidence exists for coronary heart disease and aging [24, 18]. There is even some evidence that SRH predicts the prognosis after percutaneous coronary stenting [25].

Regarding the association between smoking and SRH strong evidence exists. An Irish study showed that nonworking status, no private health insurance, inability to afford enough food, no car, being non-married, low social participation, serious neighbourhood problems, low social support, smoking, no alcohol consumption, illicit drug use, low physical activity and obesity were associated with poor SRH [26].

Only little is known about the influence of language on the SRH. A study in Singapore showed that Chinese speaking participants reported a lower SRH than English speaking participants despite that only very few differences existed between both groups, except the language [14]. In our study Dutch-speaking participants reported a lower SRH than French-speaking participants. The importance of language as such is unclear, because important cultural and socio-economic differences exist between both groups.

Our study detected an association between low SRH and not knowing one's own length. Although only few participants didn't know their length, we found it to be a strong predictor of low SRH. To our knowledge no information exists on this association, but other studies do mention an association between knowing one's own medical condition and SRH [15]. It is remarkable that in our study this association was only found for length and not for blood pressure, weight or other evident parameters. Having a family history of breast cancer or coronary heart disease also seems to impact on SRH. Knowing that one has an increased risk for a hereditary disease seems to affect one's SRH. However, this was not confirmed for colon cancer.

In our study, being on a cholesterol-lowering diet was associated with a high SRH, confirming previous evidence that people who are on a diet report a higher SRH [27]. One might expect that the quality of life of people on a diet is lower than of those who are not. Leading a healthy lifestyle probably has a favourable impact on the SRH. However, people following a diet often have an unhealthy lifestyle and compensate with a diet.

Conclusions

Our findings on the associations between low SRH and diabetes, increased body mass index, coronary heart disease, smoking and aging are known and confirm what has previously been reported in literature. However, the associations between low SRH and speaking Dutch, not knowing your own length or having a family history of breast cancer or colon cancer as well as the association between high SRH and being on a cholesterol-lowering diet are new and interesting findings. This is important because these factors might have an impact on the assessment of SRH.