Characteristics of the Respondents
Table 1 presents summary statistics for our sampledisaggregated by region of residence (rural/urban). We report four statistics: mean, the standard deviation, minimum and maximum value for each variable. In the final column, we report the p-value to test the equality of the means between the rural and urban samples.
Firstly, we look at the demographic characteristics. Our sample is evenly split between rural and urban areas (48% are urban residents) and about half are men. A typical respondent is about age 50 (the youngest is 14 and the oldest is 71), who lives in a household with 3 members and self-reported earning an annual income of 86,000 yuan (the median is 50,000 yuan and the highest is 3,500,000 yuan). In terms of education, about 9% are illiterate, 16% finished high school, and 17% have a degree of vocational or above. In terms of occupation, about 9% are working in public sectors (including civil servants, medical workers, and teachers), 29% are farmers, 18% are manual labourers, and 17% are working in private sectors. As mentioned earlier, Ningbo is a one of the well-developed coastal cities in China, thus the social-economic status of residents is above the national average.
We now turn to their health literacy and conditions of chronic diseases. Table 1 shows that the level of health literacy on CDP is 25.7%, meaning out of 100 people living in Ningbo, 26would give correct answers to more than 80% of questions and be considered as having health literacy on CDP. The prevalence ratefor chronic disease is 26%. The most prevalent disease type is hypertension (19%) followed by diabetes (5%) and heart problems (2%). The prevalence rate for cerebrovascular diseases and cancer is not high, about 1%, respectively.[1]
Significant differences also arise in rural and urban samples in terms of health literacy and chronic diseases. The urban residents have a higher level of health literacy on CDP. At the same time, they have fewer chronic diseases.Urban residents are also significantly younger (46 vs 51 years), which we think is partly due to the rural-urban migration, whereyounger people from rural areas come to urban areas for better job opportunities. Urban residents tend to live with fewer household members (2.8 vs 2.7). They earn more (102,241 vs 71,195 yuan) and are better educated (19% vs 49% in terms of the proportion of an education of high school or above). Not surprisingly, they are also more likely to work in public sectors and less likely to work as a farmer.
Characteristics of Groups with Different Level of Health Literacy
From Table 1, we find urban residents are significantly better-off in many aspects: they are healthierand have a higher level of health literacy on CDP; and they are younger, better educated and wealthier. In order to investigate the relationship between chronic diseases and health literacy, we further group our respondents by their status of health literacy on CDP(hereafter we term them ‘health literate’ or ‘health illiterate’ in short) in Table 2 to examine their respective characteristics.
Not surprisingly, we find the prevalence of chronic diseases is significantly lower among the ‘health literate’ group. In addition, the‘health literate’groupare more likely to live in the urban areas, are younger (45 vs 50), have a higher income, are better educated, and more likely to work in public sectors or employed in private sector.[2] Similar patternsare also observed both in the rural andthe urban sample (See Table A2 in the Appendix).
While we observe a lower prevalence rate of chronic diseases among ‘health literate’residents (i.e. classified as having health literacy on CDP), we also find this‘health literate’ group younger, better educated, and wealthier, which are all factors that are associated with a lower likelihood ofhaving chronic diseases. In other words, the negative relationship we observe between health literacy and chronic disease may not reflect the causal effect that health literacy has on preventing chronic diseases, but actually reflect the observed characteristics, such as age and education in preventing chronic diseases.[3] Next, we will take into account these ‘confounders’ to untangle the relationship between health literacy and chronic diseases.
Does Health Literacy Prevent Chronic Diseases?
We predict the occurrence of chronic disease with a set of hierarchical equations in Table 3. In column (1) we include no covariate but the binary variable of health literacy alone.Actually, the estimate of in column (1) will merely replay what we observed in Table 2. In columns (2)-(4), we add sequentially three blocks of variables to the equations, representing, in order of entry, region of residence, gender, income and household size; occupation; age and education. This ordering provided a means to observe how each block of variables added in later explained the effect of health literacy shown in column (1).We will show age and education are the main confounder that explained away the effect of health literacy in column (1). In column (5) we included the full set of covariates rendering us a ‘purer’ effect of health literacy, which partials out potential confounders.
The first equation in column (1) reveals that having ‘health literacy on CDP’ is associated with a reduction in the likelihood of having chronic disease by 4.8 percentage points. The second equation in column (2), which added gender, annual income and number of household members, shows that higher income is also associated with a lower likelihood of having chronic diseases and the effect of health literacy remains negative despite a small reduction in magnitude. The effect of household size is also significant, showing that respondents living in a larger household are less likely to have chronic diseases. Equation three in column (3) shows that occupation is also a strong predictor of the respondent’s chronic disease condition. Compared to those working in public sectors, farmers (manual labourers) have a higher probability of having chronic disease by 25 percentage points (11 percentage points). More importantly, with the inclusion of occupation, is now half the size as before, implying occupation explains away part of the negative effect health literacy has on chronic diseases. In column (4), we include age and education. The estimate of changes sign and is significant at 10% significance level, implying having health literacy ‘increases’ rather than ‘decreases’ the likelihood of having chronic diseases. The size of this effect is not negligible, about 2.3 percentage points. The effects of age and education are expected. Those who are younger and better educated are less likely to have chronic diseases. Those effects are significant both statistically and economically, showing they are important predictors of having chronic diseases.
We thinkage and education are the main confounders to the relationship between health literacy and chronic diseases we observe in column (1)and we are provided with this clue in two places. Firstly, there is asubstantial increase in R-squaredin column (4) at the bottom of the table compared to columns (1)-(3). Secondly, in column (5) we include the full set of covariates and the estimate of health literacy is unaltered compared to column (4). Similar patterns of results are also observedin split rural and urbansamples (TablesA3 and A4 in the Appendix).[4]
How can we explain this positive association between having health literacy and chronic diseases occurrence once we have controlled for age and education? We have to take into account the prognosis of chronic diseases and its interaction with acquiring knowledge about diseases. It is possible that people acquire the knowledge about the diseases (thus becomes ‘health literate’) after they havehad the disease. Other than books, newspapers or magazines,people can access health knowledge from doctors[10]. Therefore, although the estimate is positive, it does not mean having health literacy is bad, but having chronic diseasemight help a respondent toacquire health literacy on CDP. If this is the case, we are likely to find a stronger effect among the elderly, who are more vulnerable to chronic diseases. We conducted this exercise bysplitting the sample by age and we report the results in Table 4 (next page).Indeed, we find the positive association between health literacy and chronic disease is only present among those aged 60+ but is absent in the two younger age groups. Similar results are found for rural and urban sample in TablesA5 and A6 in the Appendix.
Does Having Chronic Disease Help to Acquire Health Literacy?
Despite Table 4, we are unsure whether this indeed is the case unless we explicitly estimate a model that predicts the probability that a respondent has ‘health literacy on CDP’. This is our task in this section where we predict the probability that a respondent has ‘health literacy on CDP’ using a linear probability model. As before, we estimate the model using the full sample and then rural and urban samplesseparately. Because differing results arise in rural and urban samples, we report only urban and rural results in Table 5.We discuss firstthe urban results as a benchmark in Panel A and then highlight differences in rural results in Panel B.
Controlling a series of characteristics of the respondents (gender, annual income, household size, occupation, age and education), we find those with at least one type of chronic disease are indeed significantly more likely to be ‘health literate’ by 3 percentage points (column 1). If health literacy can be acquired in response to a negative shock such as chronic diseases, we are more likely to observe this effect among those who bear the shock not long ago than those who had itlong time ago. We do not have the retrospective data on the change of health literacy during the prognosis of the disease on a respondent, but we could compare the level of health literacy among those whose first chronic disease was diagnosedless than one year ago and those with their first disease longer. This is what we did in our second equation reported in column (2). It shows that among the group whose first chronic disease was diagnosed within the last year, there is a boost in acquiring health literacy compared to those without chronic diseases, but this effect is absent among those whose first chronic disease was diagnosed 2-4 years ago or earlier. Besides, it appears having more than onedisease(that is comorbidities) increases the likelihood of acquiring health literacythan having only one disease as shown in column (3), but this difference is not statistically significant.[5]
Next, we examine whether this relationship is related to specific type of disease(s). This is done by replacing the number of diseases with six dummy variables indicating the types of diseases in column (4). We find having hypertension is likely to contribute to acquiring health literacy by 4 percentage points (that is 14% increase over 25.8 percentage points - the base rate of health literacy level). It is worth noting the effect of having cancer.This estimate is insignificant but sizable (not very surprising given the prevalence rate for cancer is less than 1% in the sample), showing those with cancer have a higher probability of acquiring health literacy on CDP by 10 percentage points. This sounds ironic but not counterintuitive. More importantly, this finding lends support to our speculation that negative shock prompts people to respond. Thus here, the affected patients (by chronic diseases) are more likely to acquire health literacy on CDPceteris paribusthan those who were not affected.
Compared with the results for the urban sample, the effects of duration and the types of diseases in columns (2) and (4) differ significantly from the urban results (see Panel B in Table 5). For rural respondents, those whose first chronic disease was diagnosed more than 5 years ago are significantly less likely to have health literacy on CDP than the healthy people (not having any chronic diseases). In rural sample, having heart problems is the only disease type that is significantly associated with having health literacy on CDP.
Several potential hypotheses stem from the juxtaposition of these results. First, diagnosis with chronic diseases helps an individual to acquire health literacy on CDP. Second, the acquirement via this channel however is more likely to occur when the respondents were diagnosedwith the disease not long ago.Finally, the response to negative health shock also differ by disease type. Acquiring health literacy is more likely to occur when the respondent wasdiagnosed with hypertension or cancerfor the urban resident (or heart problems for the rural residents). Compared to other chronic diseases, the diagnosis of hypertension is reasonably inexpensive and accurate [14] andthe relationship between hypertension and several other diseases, such as cerebrovascular disease, heart diseases has been widely accepted. The implication is an early discoverymight be helpful.
The effects of other variables have expected signs, which are reported in TableA7 in the Appendix.For example, those who work in public sectors are more likely to have health literacy than farmers; older respondents are less likely to have health literacy (but it is only significant in rural areas) and higher education is associated with higher likelihood of having health literacy on CDP. In particular, for the urban sample, we find a positive association between household size and the probability of having health literacy on CDP, but not in rural sample. We think this might arise because in a larger household, an individualis more likely to be supported by family members especially when there are young members, who have a stronger incentive to acquire new information because the payoff period for any information investment is longer for them [15].
Acquiring Health Literacy after Chronic Diseases: Can Health Literacy Prevent Comorbidity?
Now we are back to the question we asked at the outset, but in a slightly different form. If being diagnosed with achronic disease also helps one to acquire health literacy on CDP, could this acquirement prevent the respondent from anewchronic disease? That is to say, does health literacy help patients deal with a comorbidity before it takes place? For example, we might be interested in knowing whether having health literacy reduces the likelihood of having another disease such as hypertensionfor patients diagnosed with diabetes. We will address this question using the following specification:
where indicates the hypertension condition of a respondent (=1 if has hypertension; =0 otherwise); indicates the diabetes condition of a respondenti (=1 if has diabetes, =0 otherwise). and have the same definitions as before. The estimate of informs us in the absence of diabetes, of the effect that health literacy has on an individual’s likelihoodof having hypertension. The estimate of is likely to be positive because chronic diseases are often caused by common risk factors, thus, having one chronic disease is usually associated with having another chronic disease. Our key estimate of interest is . If it is negative, it implies that the effect that health literacy has on hypertensionoccurrence changes withwhether the respondent also has diabetes; health literacy fills a gap among patients with one condition to help them to prevent a new condition.[6] Strictly speaking, we could not interpret acquiring health literacy as occurring after the diagnosis of diabetes because our data is cross-sectional.
We experimented the above specification alternating the predicting disease variable and the explanatory disease pairs. And we do it for three samples, full sample, rural sample and urban sample, separately. We find among urban samples, there are five pairs of disease types that entail a significant interaction effect but not for the rural or the full sample and we report it in Table 6. Separate results for rural sample are available upon request.
In columns (1)-(2), we predict the probability of having comorbid cerebrovascular diseases. Expectedly, having heart problems raises the likelihood of cerebrovascular disease by 6 percentage points when an individual does not have health literacy on CDP. The coefficient on health literacy is not significantly different from zero, meaning health literacy has little role to play in preventing an individual from having cerebrovascular diseasesas the first chronic disease. However, if an individual has had heart problems,having health literacy reduces the likelihood of having cerebrovascular disease by 7 percentage points. This interaction effectcould more than offset the comorbid effect of having heart problems. In column (2), we replace health problems withcancer and again predict the probability of having comorbid cerebrovascular diseases. Having cancer is associated with a higher probability of having cerebrovascular diseases (by 5 percentage points) and the interaction effect is 6 percentage at borderline significance, which again could more than compensate the positive comorbid disease effect.[7]
In columns (3), we predict the probability of having comorbid heart problems with cerebrovascular disease (the reversed case as in column 1). Having cerebrovascular disease is strongly associated with a respondent’s likelihood of having heart problems when the respondent has no health literacy on CDP. The size of interaction effect is substantial. If a respondent has had cerebrovascular disease, health literacy on CDP could help prevent the respondent from having heart problems by 23.4 percentage points.
In columns (4)-(5), we predict the probability of having comorbid diabetes. The interaction effect is insignificant but sizable, showing health literacy reduces the likelihood of having diabetes by 4 percentage points if a respondent has heart problems. Similarly, health literacy reduces the likelihood of having diabetes by 16.4 percentage points if a respondent has cerebrovascular diseases.
Sensitivity Analyses
In this section, we look into the sensitivity of our results. We added regional fixed effects (12 dummies indicating monitor stations or 112 dummies indicating village/communities) and re-estimated equation(2). The results are not altered with the inclusion of regional fixed effects (reported in Table A8 in the Appendix). Similar to what we have in Table 6: the interaction effects become greater in sizebut the significance is not altered, showing our findings are not confounded by the heterogeneity of respondents coming from different monitoringstations or different village/communities.
Next, we apply the sample weights and re-estimated equations (2) (reported in Table A9 in the Appendix). A noticeable difference is the interaction for cancer reduces in size and significance but all else are similar, and heart problems and cerebrovascular diseases in columns (1) and (3) remain significant.
Although LPM is easier to interpret, they might suffer from problems such as the error terms will not be normally distributed, there will be heteroskedasticity, and predicted values will fall outside the logical boundaries of 0 and 1. We re-estimated parallel results for Table 3 and 5 using logit model and find similar results(reported in Table A10 and Table A11).[8]
We find that using specific health literacy is relevant to our studying the occurrence of chronic diseases, we also experimented using alternative health literacy variables given the availability of a more generic measurement of health literacy.Our key information is not changed. These results are not reported but available upon request.