The association between voluntary work and health care use among older adults in Germany

Sample

Data for this analysis were derived from the German Ageing Survey (DEAS – “Deutsches Alterssurvey”) which was conducted as a nationwide representative cross-sectional and longitudinal study of the community-dwelling population aged older than 40 years in Germany. The survey is funded by the German Federal Ministry for Family Affairs, Senior Citizens, Women, and Youth (BMFSFJ). Since 2002, the study is organized by the German Centre of Gerontology (DZA – “Deutsches Zentrum für Altersfragen”). The cross-sectional baseline samples have been set up on a recurring four-year basis starting in 1996 and ending in 2014. These baseline samples were stratified by gender, age as well as region (Eastern/Western Germany) and randomly selected from population registers within municipalities. If a written permission was given, participants from the baseline samples were re-contacted to take part in additional waves. Data collection for panel assessments were performed in 2002, 2008, 2011, 2014. The sixth wave (2017) is currently conducted.

The objective of the DEAS is to provide a unique database that enables researchers to study the multifaceted living conditions of the German population aged older than 40 years as well as the consequences of aging on individuals and processes of social transition. Therefore, participants were asked questions regarding their socio-demographic circumstances, their living conditions as well as different topics related to aging by using oral face-to-face interviews. These personal interviews took place at the participants’ homes and were carried out by trained interviewers organized as Computer-Assisted Personal Interview (CAPI). At the end of the interview, respondents were asked to fill out a standardized questionnaire that covered, for example, subjective attitudes as well as psychological topics and health.

Due to data availability in the outcome measures and our variable of interest, we restricted the analyses to the second (2002), third (2008) and fourth (2011) waves. Differences between the sample sizes were due to different recruitment approaches. In 2002, 2008 and 2011, individuals who had participated in previous waves were asked for a re-interview. New study participants were solely included in 2002 and 2008. The overall sample size in 2002 was 5194 individuals (of whom 3670 were new and 1524 were re-interviewed) and in 2008 8200 individuals (of whom 6205 were new and 1995 were re-interviewed). In 2011, 4854 individuals took part in the survey interviews (of whom 0 were new and 4854 were re-interviewed). Response rates were 38% in 2002, 38% in 2008 and 56% in 2011. In total, the response rates of the DEAS were consistent with other German surveys, but rather low compared to other European studies on aging [25]. Time or health restrictions as well as refused re-participation were the most common reasons for missing follow-up data [26]. More details of the DEAS have been reported elsewhere [27]. The study was conducted in accordance with the ethical standards of the Helsinki Declaration. Small incentives were provided to all individuals who participated.

Variables

Statistical analyses

To estimate the impact of volunteering in groups and organizations on the use of health care services, FE regressions were used. As a special feature, FE regression models allow for the association between time-constant factors and the explanatory variables. Under the assumption of strict exogeneity, FE regressions lead to consistent estimates [34]. In contrast, techniques such as the pooled ordinary least square (POLS) or the random effects (RE) would lead to inconsistent estimates when unobserved factors and the explanatory variables are correlated. Time-constant unobserved factors, such as gender or genetic disposition, are a widespread topic, especially in social sciences that need to be considered within modern research [35].

Our choice towards the FE specification has been confirmed by performing the Hausman test [36]. It basically investigates whether there is an association between the unobserved time-constant factors and the explanatory variables. The null hypotheses (they are not associated) were rejected for all outcome measures. Therefore, FE regressions were favored against RE regression models and used within this analysis (see Additional file 2).

By removing all time-constant factors, FE estimates are based solely on changes within individuals over time (intra-individual changes). As a consequence, the FE estimator is not biased by time-constant unobserved heterogeneity. It is also called “within-estimator” and enables to estimate causal effects by comparing intra-individual changes (with certain restrictions). Since all between-unit variation is eliminated, time-constant variables cannot be estimated in FE regression analysis. Nevertheless, these variables can be used for descriptive purposes and in terms of moderator variables in sensitivity analysis [35].

To estimate the predictors of GP and specialist visits, we conducted a FE Poisson regression, which is a commonly used model for measuring count-data [34]. As suggested by Stock and Watson [37], cluster-robust standard errors (SE) were used in order to avoid substantial underestimation of the true standard errors. To estimate the predictors of the binary outcome variable hospitalization, a conditional FE logistic regression was applied [38].

For sensitivity analysis, the main model was extended by including an interaction term composed of level of educational status and volunteering (volunteering x education). The underlying idea was that the impact of volunteering on the use of health care services might differ by educational level [39]. As there is evidence that in Germany formal volunteering is more likely performed by men [40], we further tested whether gender-specific links between volunteering and the outcome variables exist. Therefore, we included a respective interaction term (volunteering x gender).

The proportion of missing values was smaller than 3% for all explanatory variables, except for the variable income which had less than 6% missing values. To achieve statistical significance, explanatory variables need to reach a p-value smaller than 0.05. Statistical analyses were conducted using Stata 14 [41].

Sample characteristics

The pooled (2002, 2008 and 2011) median of GP visits and of specialist visits was 2.5 and 5.5 consultations in the 12 months preceding the interview. In this time span, 88.9% of the individuals consulted a GP and 95.3% a specialist at least once. Moreover, about 18.6% spent one or more nights in the hospital during the year preceding the interview.

FE regression with GP visits as outcome measure was based on a total number of n = 6586 participants. Regarding the time-constant variables, which are not included in FE regressions, 48.4% of the participants were female and most participants had a medium level of education (51.9%). Regarding these participants which are included in FE regression analyses, mean age was 63 years (± 11 years), with a range of 40 to 95 years. More than half of the sample consisted of retired individuals (51.7%), was married, living together with a partner/spouse (74.9%) and reported adequate accessibility of doctors and pharmacies (82.8%). Furthermore, mean monthly equivalent net income was €1728.7 (± €1436.1). Most of the individuals had overweight (42.9%) or normal weight (36.9%). The mean self-rated health was 2.4 (± 0.8) and the mean number of chronic diseases was 2.4 (± 1.8). Only 16.5% were currently smoking. As for volunteering, 21.1% of the individuals reported to be involved in volunteering.

In total, descriptive characteristics for participants included in FE regression analysis with specialist visits as outcome measure were similar. For individuals with hospital stay as outcome variable descriptive statistics differed slightly. For example, mean age was 65.1 years (± 10.8 years) and the mean number of chronic illnesses was 2.9 (± 1.9). Concerning the voluntary work, 19.2% of participants volunteered (21.3% of individuals with specialist visits as outcome measure). However, the total number of observations mainly differed among the three outcome variables due to the different number of observations with intra-individual changes over time in the outcome variables. Thus, FE regression analysis with specialist visits as outcome measure was based on a total of n = 7068 individuals, and was based on a total of n = 2186 individuals with hospital stays as outcome measure.

To avoid imprecise regression estimates resulting from time-varying variables with only little within variation [42], the variables were tested before including them in the FE regressions. However, our data showed sufficient within-unit variation in order to include all variables in the FE regression models.

Correlations

To get a more comprehensive picture of our data, pairwise cross-sectional correlations were computed. To account for the problem of multiple comparisons, Bonferroni-adjusted significance levels were used. Detailed results are reported in Additional file 3. Regarding volunteering and the three outcome measures, we found that volunteering was negatively correlated with GP visits (r = − 0.06, p < 0.001), whereas it was not significantly associated with the probability of hospitalization and specialist visits. Among the other explanatory variables, the highest correlation was found between self-rated health and chronic diseases (r = 0.44, p < 0.001). There was no correlation value higher than +/− 0.50 which indicates that the correlation level of the variables within our analyses can be assessed as low [43].

Regression analysis

For specialist visits, results of the FE poisson regression indicated that specialist visits increased significantly with changes from “no volunteer involvement” to “volunteer involvement” (β = 0.06, p < 0.05), after adjusting for control variables. Among the predisposing factors, it was shown that the number of specialist visits decreased with higher age (β = − 0.01, p < 0.001). Concerning enabling factors, none of the variables affected specialist visits on the statistical significance level. Among the need factors, the number of specialist visits increased with decreased self-rated health (β = 0.14, p < 0.001). Moreover, specialist visits increased with the number of chronic diseases (β = 0.06, p < 0.001).

For sensitivity analysis, a volunteering gender interaction term as well as a volunteering education interaction term was included. The added interaction terms did not identify any significant effects. Additional file 4 depicts detailed results of the sensitivity analysis with specialist visits as outcome measure since solely for specialist visits a significant effect has been found in the main model.

Main findings

In the present study, we used data from a representative sample of adults older than 40 years in Germany to investigate whether volunteering is associated with health care utilization, covering GP and specialist visits as well as nights in the hospital longitudinally. Pairwise correlations revealed that volunteering is correlated with GP visits. After controlling for possible confounders, the results of our retrospective analyses indicate that the number of specialist visits increased with changes in volunteering, whereas GP visits and hospitalization were not affected by volunteering. Besides, all three outcome measures were significantly associated with self-rated health. For sensitivity analysis, we tested whether gender or educational level moderate the effect of volunteering on health care utilization. None of the interaction terms did achieve statistical significance.

Relation to previous research

Since the majority of the existing literature is based on physical and mental health-related outcomes rather than patterns of health care utilization, it is difficult to compare our results with preceding studies. To the best of our knowledge, Kim and Konrath [22] were the first researchers who analyzed the association between volunteering and health care utilization. After controlling for a number of covariates, they found that volunteers have a greater probability to use preventive health care services and spend a lower number of nights in the hospital compared to those who do not volunteer. They did not find a significant association between volunteering and the frequency of doctor visits. Compared to our study, findings were partly different. This might be due to differences in the design and structure of the studies. By analyzing intra-individual changes over time, our study extends current knowledge [22]. The existing study focused on models that compare individuals who volunteer with non-volunteers. This might lead to an overestimation of the impact of volunteering on health care utilization since they did not control for (time-constant) unobserved heterogeneity. Thus, drawing conclusions about causal effects might be difficult. Moreover, they did not distinguish between different medical specialties, even if the predictors are very likely to be influenced by that.

There are a number of studies looking at more general concepts of volunteering, such as social relationships or social interactions that aimed at analyzing an association with health care use [44, 45]. It might be that these concepts are somewhat related to volunteering. As they are not completely congruent with formal volunteering itself, we refrained from comparing them with our findings.

Initially, we hypothesized that volunteering is negatively associated with health care utilization because of the positive health related outcomes linked with the volunteer involvement of older adults. We could not confirm this hypothesis. There are several possible explanations for more visits to specialists resulting from changes in the involvement in volunteering. For example, people who volunteer are characterized by having more empathic traits than non-volunteers [46, 47]. Consequently, volunteers might be more concerned about how their poor health status affects the well-being of others and therefore, they are interested in staying healthy. This in turn could lead to an increase in the use of health treatments, such as preventive services. This explanation is in line with Kim and Konrath [22]. However, this explanation is based on the assumption that preventive services are often performed by specialists, e.g., cervix screening at gynecologists [48]. Another possible explanation is that volunteering appears to ease access to health-related information [13, 49]. This might alert people to behave more health-conscious and could in turn lead to a higher amount of specialist visits, again due to increases in preventive services. Thereby, it is likely that people gather health-related information particularly through voluntary activities which are related to areas of health (e.g., volunteering in hospitals). Additionally, volunteers are not only concerned about others but are likely to have a strong sense of their own value. For example, Okun [50] found that volunteering is related with a higher level of self-esteem. People who volunteer might care about their health by using more preventive services. This might be a possible explanation of the higher rate of specialist consultations of the volunteers. Moreover, being involved in volunteering can lead to psychological stress and burnout. Scientists found that volunteering within onerous areas, such as volunteering in hospices, can cause psychological stress which may lead to the abandonment of the voluntary work [51]. Thus, the increase of specialist visits could result from more consultations with, e.g., neurologists. According to previous research, volunteering was found to be positively associated with physical activities [13]. The higher amount of specialist visits might stem from injuries or accidents resulting from those physical activities (e.g., orthopedists). Moreover, it was shown that the number of contacts with specialists increases significantly with higher educational level [52]. As we already mentioned, there is a positive association between the level of education and the probability to volunteer [39], which supports our findings.

However, as there is evidence that volunteering is associated with better physical and mental health, it is somewhat surprising that no statistically significant results were found for GP visits nor hospitalization in the FE regression analyses. It is difficult to explain why solely associations with the number of specialist visits were significant but not with GP visits considering that people generally contact GPs first as they often operate as gatekeeper and coordinators [44]. Free accessibility to physicians of all medical specialties, as it is usually the case in Germany, might explain these findings. Additionally, scientists found that people suffering from multi-morbidities contact more medical specialties [53]. As our data set consists of older adults who are generally more likely to have complex diseases and multi-morbidities [54], this could confirm our explanation. It might be that these diseases are sufficiently treated by specialists and therefore, hospitalization is not required. However, it is worth noticing that these explanations are only of theoretical nature and based on the assumption that serious illnesses generally require the involvement of (several) specialists.

Strengths and limitations

This is the first study examining the impact of volunteering on the use of health care services in Germany. Data were derived from a large, population-based study of community-dwelling older individuals. By using panel data methods (FE regressions), we provide a means of controlling the impact of time-constant unobserved factors. Another strength refers to the distinction between GPs and specialists as literature found educational and health-related differences in the use of GPs and specialists [52].

It is worth noting that the FE estimator uses only within-variation of the individuals. Thus, our findings refer to individuals with changes in the use of health care services during the given time period (average treatment effect of the treated) [35]. We cannot dismiss the possibility that estimates might be biased due to reverse causation. For example, it could be possible that after being treated by a specialist, e.g., for reasons of a complex disease, individuals seek to give something back and to be meaningfully active. For this purpose, volunteering could be a good venue. The application of panel instrumental variables methods to address the endogeneity problem, require strong assumptions [34]. Therefore, FE regressions were used in this study. It is important to take potential sample selection bias and panel attrition into account. However, both effects were shown to be small in the German Ageing Survey [27] and it has been demonstrated that panel attrition is not necessarily an issue when selected longitudinal aging studies examine the association between variables [55, 56]. Moreover, we cannot dismiss the possibility that other time-varying factors (e.g., changes in accessibility) exist that bias our estimates. Future research is required to clarify this issue. Lastly, for reasons of data unavailability, our analysis was restricted to three waves with relatively long time spans between waves. As a result, short-term changes might not be recognized. Also, due to data unavailability, it was not possible to include further lifestyle factors, such as alcohol consumption or physical activities. We could neither distinguish between the specific type or intensity of the voluntary activity nor between different forms of health insurance (statutory and private health insurance) due to reasons of data availability. Future research is required to clarify these issues.