Impact of child’s migration on health status and health care utilization of older parents with chronic diseases left behind in China

Physical health were measured by two indicators: self-rated health (SRH) and physical limitations. SRH represents the own assessment of participants. CHARLS provides two dimensions of self-assessment health status, respondents were divided into two groups randomly, rating their health status from excellent to very good, good, fair, poor and from very good to good, fair, poor, very poor. In this study, SRH were coded as a dichotomous indicator, we considered SRH of good or above in both dimensions as the cutoff of good self-perceived health [33]. For regression model applicability, we redefined “excellent”, “very good” and “good” as good health status assigned a value of 1 and “fair”, “poor” and “very poor” as bad health status assigned a value of 0. The CHARLS questionnaire asked the respondents to complete two self-health assessments at the beginning and the end of the interview respectively. Considering that some of the respondents may not be very clear about their health status at the beginning of the interview, the second answer was chosen in this study because of its higher credibility, and for respondents who did not have a second answer, we chose the first answer. The variable physical limitations were also chosen to represent physical health because such a method allowed us to gauge older adults’ difficulties in activities of daily living. Disability was measured by two scales: the 6-item scale of activities of daily living (ADL) refers to physical activities such as dressing, bathing and eating and the other 5-item scale of instrumental activities of daily living (IADL) refers to self-care ability of life such as preparing meals, shopping and taking medications [34]. The options were set to “No, I don’t have any difficulty”, “I have difficulty but can still do it”, “Yes, I have difficulty and need help”, “I cannot do it”, and corresponding score was “0 points”, “1 point”, “2 points”, “3 points”. Having no disability was defined as having no difficulty in all ADL and IADL items which means the sum of the responses of all the 11 questions equals 0 [34].

Mental health were also measured by two indicators: self-rated life satisfaction (SRLS) and the 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10) questionnaire. To assess the SRLS, respondents were asked to select from “Completely satisfied”, “Very satisfied”, “Somewhat satisfied”, “Not very satisfied” and “Not at all satisfied” to answer the question of how they think about their life-as-a-whole and how they satisfied with it in CHARLS. In this study, the first three items were defined as “Satisfied” and the last two were defined as “Not satisfied”, assigned 1 and 0 respectively. Depression scores were calculated according to the CES-D-10 which includes 10 questions. The response for each CES-D-10 question was on a four-scale metric: rarely, in some days (1–2 days per week), occasionally (3–4 days per week), and most of the time (5–7 days per week). Participants’ responses were recoded from 0 (rarely) to 3 (most of the time) for the negative questions and from 3 (rarely) to 0 (most of the time) for the positive questions (The item “I felt hopeful about the future” and “I was happy”). We added up the scores to create a depression indicator ranging from 0 to 30, and higher values suggest higher levels of depressive symptoms. The CES-D-10 indicates good validity and reliability, with a Cronbach’s alpha of 0.815 [35]. A cut-off score of ten or higher was adopted to indicate the respondents who had significant depressive symptoms in this study [36].

Health care utilization was measured based on the following questions: The binary variable of the outpatient visits in the last month was measured by the question “Have you visited a public or private hospital, public health center or clinic, health worker’s or doctor’s practice, or been visited by a health worker or doctor for outpatient care in the past month?”, respondents who did go to a health care provider were coded as 1 for utilizing outpatient care while those who did not were coded as 0 for underutilizing outpatient care; The other one is the binary variable of the inpatient visits in the last year which was measured by the question “Have you received inpatient care in the past year?”, respondents who received inpatient care were coded as 1 for utilizing inpatient care while those who did not were coded as 0 for underutilizing inpatient care. The answers to the question were “yes” or “no”.

Numerous variables are available in CHARLS. In this study, we focused on the chronically ill elderly people who had at least one child, using migration of child as a main independent variable. By enquiring “Where does this child normally live now?” with seven response options (This household, and economically dependent; This household, but economically independent; The same or adjacent dwelling/courtyard with me; Another household in my permanent address’s village/neighborhood; Another village/neighborhood in my permanent address’s county/city/district; Other: province_city_county/city/district, _village/neighborhood; Abroad), we collected information about their child’s habitation. In this study, having a migrant child was defined as having a child living in another county/city/district in the same province, in another province and abroad [5]. In other words, people who responses “Other: province_city_county/city/district, _village/neighborhood” and “Abroad” were considered to have a migrant child. This variable was divided into two categories: having migrant child coded as 1 and having no migrant child coded as 0.