Background
Self-rated health (SRH) (also known as self-assessed health or self-perceived health) refers to a single-item health measure in which individuals rate the current status of their own health on a four- or five-point scale from
excellent to
poor. It is popular for its simplicity and has been extensively studied in Western populations. A series of national and international analyses has consistently shown that SRH is a good predictor of mortality of various diseases [
1‐
3]. Furthermore, many researchers have attempted to investigate the factors related to SRH and have found that SRH is strongly associated with morbidity and disability [
4‐
6]. Economic or social factors are also its main determinants [
7‐
9].
SRH is a subjective reflection of health status, called “perceived” or “subjective” health. It has been widely studied in survey research [
3,
7,
8]. However, most of the studies concerning SRH have focused on specific age groups, sex groups, or patient groups [
10‐
12]. Studies reporting SRH among the general population are few. Relative to SRH, some indicators provide direct evidence to the health status of individuals, including previous and current diseases (diagnosed by physicians), and clinical parameters measured in the hospital; these have been termed as “actual” or “objective” health. Some health-related factors, such as demographic characteristics, health history, life habit, life stress and work strain, are closely associated with objective health.
There is a complex relationship between subjective and objective health [
13]. Some studies reported that old people perceived their health in positive terms and tended to over-estimate their health [
14‐
16], while other studies found that old people tended to report poorer health than others who were with similar objective health conditions [
15,
17,
18]. Is subjective health consistent with objective health? Can SRH serve as a global measure of health status in the general population? The answers are still debatable. Thus, the assessment of the relationship between SRH and objective health status is important, for this determines whether SRH can serve as an indicator of objective health status.
Methods
This study examined the association between SRH and objective health status through a comprehensive collection of data on disease prevalence, laboratory parameters, and health-related factors in a randomly selected sample from five cities of China: Shanghai, Beijing, Xi’an, Wuhan and Guangzhou. We attempted to determine the relationship between SRH and objective health status in this population sample and to identify the factors that best discriminate the different SRH groups.
Study design and sample
We used a part of the data from our previous epidemiological survey on gastrointestinal diseases in five cities of China. The details of the methodology of the survey have been published elsewhere [
19], and are briefly described here. The survey was administered to 18,000 residents aged 18–80 years (3,600 in each city) from April 2007 to January 2008. Eligibility criteria were based on age and the mental and physical ability of subjects to participate in the survey. Residents who were illiterate, not in the 18–80 age group, or suffering from psychiatric illnesses or other disabilities were excluded from the study. They were sampled using a randomized, stratified, multiple-stage sampling method, with the age/gender distribution of the sample in accord with the distribution of local population according to the population census statistics published by the government. Thus, the sample would not be affected by the original impetus of the survey. In the first stage, one or more districts from the urban stratum and one or more counties from the rural stratum were randomly selected from each region. In the second stage, one or more blocks from the urban districts and one or more townships from the rural counties were randomly selected. In the third stage, one or more residential areas from the urban blocks and one or more villages from the rural townships were randomly sampled. Questionnaires were self-completed, with trained interviewers giving explanation on any unclear questions. The entire procedure was overseen by the supervisors to ensure the quality of the survey.
Measurements
Demographic information were gathered on gender, age, weight, height, marital status (married, unmarried, divorced, or widowed), educational level (less than primary school, completed only primary school, completed secondary school, completed high school, completed college/pre-university/university, master’s degree, or doctoral degree), total family monthly income (<RMB 2,000, RMB 2,000–4,999, RMB 5,000–9,999, or ≥RMB 10,000), etc. We used a single-item measure of subjective health—“In general, how would you rate your current health status?”—to assess self-ratings of health. Five options, which were recommended by the World Health Organization (WHO) [
20] and the European Network for the Calculation of Health Expectancies (Euro-PEVES) 2 group [
21], were listed: “very good,” “good,” “fair,” “bad,” or “very bad.”
To assess the relationship between SRH and objective health status, subjects were asked to report whether they had been diagnosed with chronic diseases, including hypertension, cerebrovascular disorder, diabetes mellitus, and chronic bronchitis. Blood sampling was carried out after overnight fasting at the Second Military Medical University Changhai Hospital in Shanghai, but this was not done in the other four cities. The following parameters were measured to assess the subjects’ health status: red blood cell (RBC), hemoglobin (HGB), aspartate aminotransferase (AST), total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), and fasting plasma glucose (FPG). RBC and HGB indicate the body’s ability to transport oxygen to tissues. AST and TC are indicators that mainly reflect hepatic dysfunction. High levels of TG and LDL-C are two major risk factors for cardiovascular and cerebral diseases. Plasma glucose provides the energy cells require and must be maintained at a certain level in order to sustain the body’s needs. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters, and the WHO criterion for Asian populations was used to classify subjects as obese, overweight, normal weight, and underweight [
22]. Data on some health-related factors were also collected, including body weight five years ago, tobacco use, frequency of physical activity, life stress, work strain, spiritual status, and quality of interpersonal relationships. Tobacco use was measured with a single question: “Do you currently smoke?” Six options were given: “no,” “1–5 cigarettes/day,” “6–10 cigarettes/day,” “11–15 cigarettes/day,” “16–20 cigarettes/day,” or “≥21 cigarettes/day.” Frequency of physical activity was measured with the following question: “How often do you engage in physical activities (e.g., physically active work, walking, riding a bike, and doing exercises)?” Four options were given: “never,” “<4 times per month,” “at least once a week,” or “at least once a day.” Life stress was measured by asking “Overall, do you feel stress in your daily life?” and work strain was measured by asking “Overall, do you feel stress in your daily work?” Life stress and work strain were assessed on a five-point scale: “no,” “a little,” “moderate,” “quite a lot,” and “extreme.” Spiritual status was measured by asking “Overall, how would you rate your spirituality?” and quality of interpersonal relationships was measured by asking “Overall, how would you rate your quality of interpersonal relationships?” Spiritual status and quality of interpersonal relationships were assessed by “very good,” “good,” “moderate,” “bad,” or “very bad.” Women were also asked to answer their ages at menopause.
Statistical analysis
All the data were doubly input into the database by two independent professional data processors using EpiData 3.1. Statistical Analysis System (SAS) 9.1.3 was used for analyzing the survey data. Data are expressed as mean±standard deviation (SD). For statistical comparison, the differences among the “healthy” (very good plus good), “relatively healthy” (fair), and “unhealthy” (bad plus very bad) groups were examined, because it was difficult to determine whether “fair” should be defined as “good health” or “poor health,” and because only a few subjects rated their health as very good (1,770 subjects) or very bad (58 subjects). Statistical significance for laboratory parameters among different groups was determined by using an analysis of variance with Bonferroni’s multiple-comparison tests. The odds ratios (ORs) and 95% confidence intervals (CIs) for the “relatively healthy” and “unhealthy” groups were calculated referenced by the “healthy” group, using logistic regression analysis by controlling for gender, age, region, marital status, educational level, current work, family income, current smoking, current drinking, frequency of physical activity, and chronic diseases. All hypothesis tests used two-sided tests and a p-value less than 0.05 was considered statistically significant.
Ethics
All the subjects had been informed that their records might be used for analysis, and had signed a written informed consent form before participation. The present study was approved by the ethics committee of Second Military Medical University.
Discussion
This study aimed to determine whether SRH could reflect objective health status and serve as a global measure of health status in the general population. For this purpose, we examined the relationships between SRH and the following: (1) prevalence of diseases diagnosed by the physicians, (2) laboratory parameters examined in the hospital, and (3) health-related factors that were regarded as risks. We found that lower SRH was associated with significant increases in the prevalence of all the diseases and abnormalities in laboratory parameters. Some health-related factors, including underweight, weight loss, marital status, education, low income, early menopause, physical activities, life stress, work strain, spiritual status, and quality of interpersonal relationships, were significantly associated with SRH. This suggests that individuals may use these factors to evaluate their overall health [
23,
24].
Notably, most subjects rated their health positively, for only 5.72% subjects reported bad or very bad health, whereas nearly 60% rated their health as very good or good. A survey conducted in Sweden among subjects aged 18–79 years showed that 7% of the men and 9% of the women had poor SRH [
25]; thus, the prevalence rate of poor SRH in our population sample, with a similar age range as theirs, was slightly lower. An obvious higher mean age was observed as the subjects rated poorer health. This means that the subjective health status of the population significantly decreases with advancing age. Similar results were found in previous studies, which showed that age was a risk factor for poor SRH [
25‐
29].
The prevalence of all the diseases included in our study, which was an important indicator of objective health, contributed to decreased SRH. This means that SRH has the ability to distinguish patients from relatively healthy people. It is noteworthy that larger ORs were found for cardio-cerebral vascular diseases, visual impairment, and mental illnesses when comparing the unhealthy group with the healthy group. An earlier study reported that with China’s rapid economic development, diseases of the heart and the cerebrovascular system have become two leading causes of deaths in this country, ranking second and third among men, and first and second among women [
30]. It is understandable that patients with cardio-cerebral vascular diseases are more likely to worry about their future, and that the diseases cause great cost and burden to themselves and their families; thus, people who suffer from these diseases rated poorer health. Visual impairment is associated with a decreased ability to perform activities of daily living and poor quality of life, particularly in the sense of depression and reduced social interaction [
31], and patients with mental illnesses had more impairment in quality of life than those with common medical disorders [
32]. These could explain why patients with visual impairment or mental illnesses rated poorer health. A previous study conducted in Sweden showed that many chronic diseases, such as neurological disease, rheumatoid arthritis, and cancer, are strongly associated with poor SRH [
29]. This conclusion has been further strengthened with the evidence from our study.
Studies examining the association of SRH with laboratory parameters were still limited. In order to further elucidate this relationship in our study, we examined the laboratory parameters in the subjects who had experienced blood sampling and listed the laboratory parameters with statistical significances. We found that better SRH was associated with increased RBC and HGB, and decreased AST, TC, TG, LDL-C, and FPG. Furthermore, subjects with abnormal laboratory parameters reported poorer health. Laboratory parameters can objectively reflect health status and are of fundamental importance for the diagnosis, prognosis and treatment of various diseases [
33]. The consistency of SRH and objective health status assessed by laboratory parameters is further evidence that SRH is a good indicator of health status.
In addition, the relationship between SRH and some health-related factors, which were regarded as risk factors for health, were examined. Obesity is surprisingly unrelated to poorer SRH, and this finding is inconsistent with many previous studies [
24,
25,
34,
35]. On further consideration, this is also explicable. In Chinese culture, being fat is not a signal of unhealthiness but a symbol of acquiring good fortune, because only the more wealthy people can afford to eat more and can thus put on more weight [
36]. This is quite different in Western countries. Underweight and weight loss are two risk factors observed in this study. Underweight or weight loss was linked to eating disorders such as anorexia nervosa, or consumptive diseases such as malignant tumors. The findings also support marriage as a protective factor for health [
37]: subjects who were divorced or widowed rated poorer health. Subjects with higher income tended to be more optimistic about their health, because they have been observed to practice more health promoting behaviors such as good dietary intake practices and physical activity [
38]. Shibuya et al. also found that low income was associated with poor SRH in a Japanese sample [
7]. Lim et al. found that marital status, education, household income, current smoking, and exercise were associated with SRH in a Singaporean sample [
39]. Life stress and work strain as risk factors for poor SRH have been examined less frequently in population studies. One study has demonstrated that family stress is a risk factor for unfavorable health-related outcomes [
40], and another has revealed that work stress is associated with poor SRH [
41]. These results support the conclusion of this study.
Our results need to be interpreted within the context of the study limitations. First, the cross-sectional nature of the study did not allow us to investigate the temporal mechanisms between SRH on the one hand and diseases or abnormalities in laboratory parameters or risk factors on the other. Second, we relied on a self-administrated questionnaire for measuring the clinical and health-related factors that were used to validate SRH. Except for laboratory parameters, all data relied on self-report. Like any other questionnaire study, the subjects’ answers may be affected by various kinds of bias such as social desirability. Finally, the study was a part of a larger study; therefore, the research questions can only be pursued within the boundaries of the original study. Some indicators were not integrated in our study due to the original focus of the study. Nevertheless, information on the majority of, but not all, possible confounders was available for analysis. There were also other factors associated with SRH such as depression, psychological well-being, health service utilization, medication usage, cognitive capacity, and social networks, but data on these factors were not collected in our study. However, with the information collected, we were able to explain the relationship between SRH and the objective health status.
The strengths of this study are as follows: (1) it includes a random sample of population from five cities of China with a broad range of ages in accord with the distribution of local population, which is a better representation of the total population. In contrast the majority of previous studies focusing on SRH mainly examined older population samples [
5,
13‐
15,
17,
18,
42,
43] or subjects with certain specific diseases [
23,
44‐
46]. (2) The present study investigates the association between SRH and objective health status from multiple dimensions, using a wide range of variables, including the prevalence of diseases, abnormalities in laboratory parameters, and health-related factors.
Competing interests
SQW, RW, YFZ, XQM, MJW, and XYY have no conflict of interests. JH has served as the director of the Department of Health Statistics, Second Military Medical University.
Authors’ contributions
JH conceived of the study and supervised all aspects of its implementation. SQW and RW assisted with the survey, completed the statistical analyses and led the writing of different versions of the manuscript. YFZ, MJW and XYY assisted with the study, and XQM assisted with the survey and data analyses. All authors contributed to conceptualize ideas, interpret findings, and review the drafts of the manuscript, and they approved the final manuscript.