- Email: [email protected]
Addictive behaviors and healthcare renunciation for economic reasons in a French population-based sample
International Journal of Drug Policy 49 (2017) 1–7
Contents lists available at ScienceDirect
International Journal of Drug Policy journal homepage: www.elsevier.com/locate/drugpo
Research paper
Addictive behaviors and healthcare renunciation for economic reasons in a French population-based sample Stéphanie Baggioa,f,* , Marc Dupuisb , Jean-Baptiste Richardc , François Beckd,e a
Life Course and Social Inequality Research Centre, University of Lausanne, Switzerland Institute of Psychology, University of Lausanne, Switzerland c Santé publique France, The French National Public Health Agency, France d Observatoire Français des Drogues et des Toxicomanies (OFDT), The French Monitoring Center for Drugs and Drug Addiction, France e ERES, Sorbonne Universités, UPMC Univ Paris 06, INSERM, Institut Pierre Louis d’Epidémiologie et de Santé Publique, UMRS, 1136 Paris, France f Division of Correctional Medicine, Geneva University Hospitals and University of Geneva, Geneva, Switzerland b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 February 2017 Received in revised form 22 April 2017 Accepted 10 July 2017 Available online xxx
Background: Healthcare renunciation for economic reasons is a major health concern, but it has been scarcely investigated among drug users, even if drug users constitute a vulnerable population in need of medical care. This study investigated associations of healthcare renunciation for economic reasons and addictive behaviors (alcohol, tobacco, cannabis, illicit drug use, and gambling) in a population-based sample of adults living in France, a country with universal health coverage. Methods: Data were collected using the 2014 Health Barometer, a French cross-sectional survey conducted among a random representative sample of the general population aged 18–64 (n = 12,852). Measures included healthcare renunciation, substance use (alcohol, tobacco, cannabis, and other illicit drugs) and gambling. Experimental/recreational and heavy/chronic use were assessed. Logistic regressions were used to test the relationship between healthcare renunciation and addictive behaviors, controlling for relevant covariates. Results: A total of 25% of the participants had renounced care at least once in the previous twelve months. Most variables of drug use were significantly associated with increased healthcare renunciation. This was the case for heavy/hazardous use and experimental/recreational use. Regular gambling was not associated with healthcare renunciation, but disordered gambling was. Conclusion: This study showed that addictive behaviors, including substance use and gambling, were part of the burden of vulnerability of people who forgo care. Therefore, drug use and gambling patterns should be a focus in the development of policies to reduce health inequalities, not only for heavy and chronic drug users. © 2017 Elsevier B.V. All rights reserved.
Keywords: Forgone care Gambling Population-based survey Substance use Health social inequalities
Introduction Healthcare renunciation for economic reasons, i.e., forgoing care due to financial reasons although clearly perceiving the need for it, is a major health concern in health inequality research. Vulnerable people are especially a focus because people in situations of deprivation and high-risk populations (e.g., those with chronic illnesses) are more likely to forgo care for economic reasons worldwide (Bodenmann et al., 2014; Guessous, Gaspoz, Theler, & Wolff, 2012; Pampel, Krueger, & Denney, 2010; Röttger,
* Corresponding author at: Institute for Social Sciences, University of Lausanne, Geopolis Building, CH-1015 Lausanne, Switzerland. E-mail addresses: [email protected] (S. Baggio), [email protected] (M. Dupuis), [email protected] (J.-B. Richard), [email protected] (F. Beck). http://dx.doi.org/10.1016/j.drugpo.2017.07.004 0955-3959/© 2017 Elsevier B.V. All rights reserved.
Blümel, Köppen, & Busse, 2016; Wolff, Gaspoz, & Guessous, 2011). Previous studies have mainly focused on the relationships of forgone care with socioeconomic status, demographics, and perceived health status. Studies found consistent associations of healthcare renunciation for economic reasons of having low income or job position, being older, being female, having dependent children, being divorced or single, and having bad health outcomes in different countries (Bodenmann et al., 2014; Guessous et al., 2012; Pampel et al., 2010; Röttger et al., 2016; Wolff et al., 2011). Another part of the literature on the utilization of health care services highlighted that drug users constitute a vulnerable population in need of medical care. In several countries, drug users have been described as having an important lack of access to primary care (Myers, 2012) and are considered medically underserved (Metsch et al., 2002), although addiction is associated
2
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
with adverse medical consequences. Studies have mostly focused on high-risk drug users’ unmet healthcare needs, such as injecting drug users (Al-Tayyib, Thiede, Burt, & Koester, 2015; Chitwood, McBride, French, & Comerford, 1999; Metsch et al., 2002; Robbins, Wenger, Lorvick, Shiboski, & Kral, 2010) and methamphetamine users (Powelson et al., 2014). This population has a large burden of unmet health needs and is more likely to visit hospital emergency departments (Vu et al., 2015). However, information on associations of occasional or recreational drug use and other addictive behaviors such as gambling with unmet health care in the general population is lacking. Some studies reported that, in the general population, unhealthy substance use was a risk factor for not receiving preventive healthcare, such as breast-cancer screening, flu vaccination, and control visits to general practitioners (Chitwood, Sanchez, Comerford, & McCoy, 2001; Lasser et al., 2011), but investigations on the relationship between forgone care and experimental or recreational drug use are needed. Overall, if healthcare renunciation is a topic of growing interest, to our knowledge, no study has focused specifically on the association between addictive behaviors and healthcare renunciation in the general population. Because drug users are likely to have unmet healthcare needs, they are susceptible to be a high-risk population for healthcare renunciation. For example, Guessous et al. (2012) reported that current smokers were more likely to renounce care than non-smokers, and Guessous et al. (2014) reported that current smokers were more likely to renounce dental care than nonsmokers. However, these studies did not report other substance use, and no information on the addictive level of tobacco use was available. This study aimed to fill this gap and investigated associations of healthcare renunciation for economic reasons and addictive behaviors (alcohol, tobacco, cannabis, other illicit drug use, and gambling) in a population-based sample of French adults. France has a universal health system, with a national public health insurance based on income. It includes significant cost sharing (approximately 30% of health care costs). Cost sharing is eliminated for individuals with several specified chronic conditions and lowincome individuals (in 2017, the upper limit for cost sharing elimination is 727s [$778] for one person, 1090s [$1167] for two persons). It is also lowered or eliminated for highly effective prescription drugs (Schoen et al., 2010). However, out-of-pocket expenditures may affect even low-income individuals (e.g., one single person having more than 727s per month). Most French citizens buy private complementary insurance that covers cost sharing (Pierre & Jusot, 2017). Data on healthcare renunciation are scarce in France. Methods Participants and procedure Data were collected using the 2014 Health Barometer, a French cross-sectional survey conducted among a random representative sample of the French population aged 15–75 years between December 2013 and May 2014. It is conducted using computerassisted telephone interviews. Telephone numbers of households (landlines and cell phones) were randomly generated, and individuals in households were randomly selected among eligible household members: having a member between 15 and 75 years old, speaking French, and living within the household during the survey. When the selected individual was unavailable, an appointment was made. Households or individuals were considered unreachable after 40 telephone calls. The response rate was 61% for landlines and 52% for cell phones. The sample included 15,635 participants, with n = 7577 for landlines and n = 8058 for cell phones. This study focused on an adult population that was 18–
64 years old (n = 13,039). We excluded participants older than 64 (65–75 years old) because they did not complete the questions related to illicit drug use. Missing values on the variables included in the study were listwise deleted (187 participants had missing values), leaving a final sample of n = 12,852. Weights were constructed in two steps: the design weight computation (which reflects the differential probabilities of selection, depending on several factors, including probability of selection of the telephone number, number of eligible individuals living in the households, number of landline and cellular telephones owned by the respondent), followed by an overall post-stratification to correct under-coverage. The questionnaire included questions on demographics, health behaviors, health status, and social determinants. The questionnaire, method, and evaluation of the method are available elsewhere (Richard et al., 2016). Measures Healthcare renunciation Participants answered whether they had renounced care for financial reasons using the following dimensions for the previous 12 months: (1) dental care, (2) eye care, (3) medical appointment, and (4) other health care. Healthcare renunciation was coded 1 if participants renounced at least to one situation in which healthcare was needed and 0 otherwise. We assessed lifetime and 12-month prevalence of substance use and gambling patterns. A 12-month period has been recommended to assess substance use behavior and problems (Dawson, 2003). Alcohol use Alcohol use was assessed using three variables: (1) alcohol use with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use; (2) binge drinking, defined as having drunk at least six drinks in a single session in the previous month (yes/no) (Herring, Berridge, & Thom, 2008), and (3) hazardous drinking (at risk/not at risk), defined usually as drinking more than 21 drinks a week for a man and 14 for a woman or drinking six drinks or more on a single occasion weekly (Mouquet & Villet, 2002). Tobacco use Tobacco use was assessed using a variable related to smoking status with five categories: no lifetime tobacco use, lifetime smoker but currently non-smoker, occasional smoker (smoking but not daily), daily smoker (with less than ten cigarettes per day), and heavy smoker (daily smoker with ten cigarettes or more per day). We classified participants who had smoked only once or twice in their lives as non-smokers. Cannabis use Cannabis use was assessed with two variables: (1) cannabis use with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use and (2) score on the cannabis abuse screening test (CAST, Legleye et al., 2015), ranging from 0 to 24, with a larger score indicating cannabis-use disorder. Previous studies reported good psychometric properties of the CAST (Legleye et al., 2013; Legleye, Piontek, Kraus, Morand, & Falissard, 2013). Other illicit drug use Participants indicated whether they used any illicit drug use other than cannabis (one question for each of the following illicit drugs: magic mushrooms, poppers, other inhalants, ecstasy,
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
3
Table 1 Descriptive statistics and bivariate associations of healthcare renunciation, addictive behaviors, and covariates. Variables
Descriptive
Healthcare renunciation No 75.0% (n = 9636)
Yes 25.0% (n = 3219)
p-value
Effect size
Age1
42.2 (13.1)
42.3
41.7
.081
–
Gender2 Male Female
47.0 53.0
79.9a 70.6b
20.1a 29.4b
<.001
.107
Living area2 Rural Urban
26.0 74.0
78.6a 73.7b
21.4a 26.3b
<.001
.049
Socioeconomic category2 Farm worker Craftsman/retailer/business owner Manual worker Employee Intermediate occupation Executive/senior-level occupation Do not know/do not answer
1.6 5.5 19.3 27.1 25.6 19.1 1.8
85.2ab 78.5ab 72.2c 68.4d 76.6a 83.1b 74.9acd
14.8ab 21.5a 27.8c 31.6d 23.4a 16.9b 25.1acd
<.001
.124
Income2 First quintile Second quintile Third quintile Forth quintile Fifth quintile Do not known/do not want to answer
14.8 17.7 17.7 22.0 23.2 4.6
60.9a 65.7b 72.6c 79.2d 88.0e 79.3d
39.1a 34.3b 27.4c 20.8d 12.0e 20.7d
<.001
.218
Subjective health status2 Good/average Bad Mental health status (0–100)1
94.6 5.4 69.5 (17.3)
76.2a 53.3b 71.7
23.8a 46.7b 63.0
<.001
.119
<.001
.047
Alcohol use2 No lifetime use Lifetime use but not current user 12-month use
3.3 7.4 89.3
64.8a 65.6a 76.1b
35.2a 34.4a 23.9b
<.001
.077
Binge drinkingb (among 12-month alcohol user, n = 11,481) No Yes (monthly or more)
80.2 19.8
76.5a 74.3b
23.5a 25.7b
.104
–
Hazardous drinking2 (among 12-month alcohol user, n = 11,481) Not at risk At risk
52.1 47.9
75.8a 76.5a
24.2a 23.5a
.417
–
Tobacco use2 No lifetime use Lifetime use but not current user Occasional smoker Daily smoker Heavy smoker (>10 cigarettes/day)
<.001
.120
65.4 9.4 6.5 18.7
78.2a 76.4ab 71.2b 65.0c
21.8a 23.6ab 28.8b 35.0c
Cannabis use2 No lifetime use Lifetime use but not current user 12-month use CAST (0–12)3 (among 12-month cannabis user, n = 1219)
57.5 32.6 9.8 2.9 (3.7)
77.9a 72.7b 65.2c 2.5
22.1a 27.3b 34.8c 3.7
<.001
.093
<.001
–
<.001
.069
.834
–
<.001
–
Illicit drug use2 No lifetime use Lifetime use but not current user 12-month use
86.9 11.0 2.1
76.1a 68.0b 64.3b
23.9a 32.0b 35.7b
Gambling2 No 12-month use 12-month use Active use (once a week or 500s/year) CPGI (0–24)3 (among 12-month gamblers, n = 7667)
40.4 41.9 17.7 0.3 (1.2)
74.8a 75.2a 74.6a 0.26
25.2a 24.8a 25.4a 0.42
CAST: Cannabis Abuse Screening Test, CPGI: Canadian Problem Gambling Index. a,b,c,d,e A same subscript letter within a column denotes that proportions did not differ; two different subscript letters denote that proportions differed at the .05 level. A Bonferroni–Holms correction was applied to keep a 5% error rate. Significant p-values were adjusted. 1 Means and standard deviations are given. t-tests for bivariate associations were computed, with partial eta squares. 2 Percentages are given. Chi-square tests with Fisher exact tests to compare columns proportions and Cramer V are given for bivariate associations. Line percentages are given for bivariate associations. No effect size was provided for non-significant associations. 3 Means and standard deviations are given. Negative binomial regressions were used for bivariate associations. No effect size is available.
4
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
amphetamines, methamphetamines, LSD, crack, coke, heroine, and GHB/GHL) with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use. Gambling Gambling was assessed using two variables: (1) gambling with three categories: no previous twelve-month use, previous twelvemonths use, and active gambling, defined as having gambled at least once a week or having spent 500s on gambling the previous year and (2) score on the Canadian Problem Gambling Index (CPGI, Ferris & Wynne, 2001), ranging from 0 to 24, with a larger score indicating disordered gambling. The CGPI has been described as a reliable instrument for non-clinical contexts (Holtgraves, 2009). Covariates Participants provided information on their age, sex, living area (urban versus rural), socioeconomic category (farm worker, manual worker, employee, intermediate occupation, executive/ senior-level professional occupation, craftsman/retailer/business owner, and do not know/do not want to answer), and household income per consumption unit. Consumption units were used to compare households of different sizes and composition by assigning a coefficient to each member of the household. We used the OECD scale to calculate the consumption units (one unit for the first adult, 0.5 unit for individuals aged 14 and older, and 0.3 for children under 14 years). Income was recoded in quintiles, with an additional category do not know/do not want to answer. Participants also evaluated their health (subjective health status), which was coded as good/average or bad. Mental health status was assessed for the previous four weeks with five questions (nervousness, discouragement, calm/relaxation, sadness, and happiness) assessing the MH-5, which is a subscale of the mental health subscales of the SF-36 with good psychometric properties (Leplège, Ecosse, Verdier, & Perneger, 1998) and computed to create a mean score ranging from 0 (bad mental health) to 100 (good mental health). Statistical analyses Descriptive statistics for healthcare renunciation, addictive behaviors, and covariates were computed using percentages and means according to the variables’ distribution. Then, bivariate associations between healthcare renunciation and all other variables were performed using t tests (continuous variables), negative binomial regressions (continuous skewed variables), and chi-square tests with pairwise comparisons using Fisher’s exact tests (categorical variables). Effect sizes were computed using partial eta squares for continuous variables and Cramer’s V for categorical variables. Then, multivariate associations of healthcare renunciation and addictive behaviors were computed, controlling for covariates. We computed logistic regressions with healthcare renunciation as the dependent variable and each addictive behavior variable as the independent variable. We reported individual coefficients and statistics on the overall goodness-offit (difference between the deviance of the model without the variable of interest and with the variable of interest). We did not perform a model including all addictive behavior-related variables together to avoid collinearity problems. A Bonferroni-Holm correction to maintain a 5% error rate was used for bivariate and multivariate associations. All analyses were performed using SPSS 24. Results Descriptive statistics and bivariate associations are reported in Table 1. A total of 25.0% of the participants had renounced care for
financial reasons at least once in the previous twelve months. More precisely, 16.4% renounced dental care, 11.1% renounced eye care, 7.1% renounced a medical appointment, and 9.5% renounced other another form of health care, with an average of 1.76 0.97 dimensions of forgoing care among those who renounced care. Demographics were associated with healthcare renunciation (first panel of Table 1): being a woman (p < .001), living in an urban area (p < .001), having a low socioeconomic status (p < .001), and having a low income (p < .001). Subjective health and mental health status were also associated with healthcare renunciation: participants who renounced care had a lower (mental) health score (p < .001). The highest effect size was for income (.218). Age was not significantly associated with healthcare renunciation (p = .081). Addictive behaviors were also associated with healthcare renunciation (second panel of Table 1). Tobacco, cannabis, and other illicit drug use were associated with increased healthcare renunciation (p < .001). Alcohol (binge drinking and hazardous drinking) was not associated with healthcare renunciation, except previous 12-month alcohol use, which was associated with decreased healthcare renunciation in comparison with lifetime and previous twelve months abstainers (p < .001). Twelve-month gambling was not associated with healthcare renunciation, but CPGI score was, with participants who renounced healthcare having a higher score of disordered gambling (p < .001). The effect sizes were small, ranging from .069 to .120 for significant associations. In multivariate associations, reported in Table 2, the results remained mostly unchanged, with substance users (tobacco, cannabis, and other illicit drugs) being more likely to renounce care, with a gradient for tobacco and cannabis. The results were also similar to bivariate associations for gambling, with a marginal increased risk of healthcare renunciation when participants had a higher CGPI score (p = .054) but not twelve-month gambling. Binge drinkers were significantly more likely to renounce care, contrary to bivariate associations that displayed non-significant associations. Twelve-month alcohol users were less likely to renounce care, but they were not significantly different from lifetime abstainers. Former drinkers, who were lifetime drinkers but had not drunk in the previous twelve months were more likely to renounce care than current alcohol users. For all models except twelve-month gambling, the model including the variable of interest was better than the model that did not include it (i.e., lower deviance). Discussion This study investigated associations of healthcare renunciation for economic reasons and addictive behaviors among a populationbased sample of people living in France with Universal Health Coverage. First, a total of 25.0% of the participants renounced care at least once in the previous twelve months, regardless of the kind of renunciation (dental care, eye care, medical appointments, and other kinds of health care). This prevalence rate is higher than those previously reported in France (15.4% among the adult population in 2008; Després, Dourgnon, Fantin, & Jusot, 2011). Forgoing care for financial reasons has increased between 2002 and 2008 (Després et al., 2011), and it seems that it has increased since then. This possible increase is worrying information. Additionally, healthcare renunciation remained a factor of social inequalities because people with low income, low socioeconomic status, and bad mental health outcomes were more likely to renounce care, as was the case in the previous decade (Després et al., 2011).
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
5
Table 2 Logistic regressions of healthcare renunciation predicted by addictive behaviors, controlling for covariates. Variables (IV)
Healthcare renunciation (DV)
p-value3
Difference in deviance4
Alcohol use2 No lifetime use Lifetime use but not current user 12-month use
0.25ab 0.27a 0.22b
.018
13.7
Binge drinking2 (among 12-month alcohol user, n = 11,481) No Yes (monthly or more)
0.22a 0.25b
.020
13.0
Hazardous drinking2 (among 12-month alcohol user, n = 11,481) Not at risk At risk
0.22a 0.24a
.063
26.9
Tobacco use2 No lifetime use Lifetime use but not current user Occasional smoker Daily smoker Heavy smoker (>10 cigarettes/day)
0.18a 0.21b 0.23bc 0.26c 0.30d
<.001
117.0
Cannabis use2 No lifetime use Lifetime use but not current user 12-month use CAST (0–12)1 (among 12-month cannabis user, n = 1219)
0.19a 0.27b 0.35c 0.49
<.001
162.6
.024
70.3
Illicit drug use No lifetime use Lifetime use but not current user 12-month use
0.22a 0.32b 0.34b
<.001
79.6
Gambling2 No 12-month use 12-month use Active use (once a week or 500 s/year) CPGI (0–24)1 (among 12-month gamblers, n = 7667)
0.23a 0.23a 0.23a 0.047
.797
0.3
.054
5.5
2
CAST: Cannabis Abuse Screening Test, CPGI: Canadian Problem Gambling Index. Dependent variable (DV): healthcare renunciation with reference category ‘no healthcare renunciation’, independent variable (IV): addictive behaviors. Separate logistic regressions were run to avoid multicollinearity, with: 1 unstandardized b for continuous IV (CAST and CPGI), and 2 adjusted means and pairwise comparisons for categorical IV, 3 p-values of Wald Chi-squares are given. 4 Differences in deviance (-2 loglikelihood) between the model without the IV of interest (but with control variables) and with the IV of interest are reported. a,b,c,d A same subscript letter within a column denotes that proportions did not differ; two different subscript letters denote that proportions differed at the .05 level. Models controlled for age, gender, socioeconomic status, income, health status, and mental health status. A Bonferroni-Holms correction was applied to keep a 5% error rate.
Addictive behaviors added to the burden of these already vulnerable individuals. Indeed, substance and hazardous substance use were associated with healthcare renunciation. Controlling for important covariates (demographics and mental health status), almost all tested variables were significantly associated with healthcare renunciation. Heavy and hazardous substance use, defined with heavy smoking (ten cigarettes per day or more), binge drinking (six or more drinks on a single occasion per month), being at risk for hazardous drinking, and having a high score on the cannabis-use disorder scale were associated with increased healthcare renunciation. Lifetime use of tobacco, cannabis, and other illicit drugs were also associated with increased forgone care. Therefore, experimental and recreational use seemed associated with healthcare renunciation and not only intensive and hazardous use, as previously shown. This result was interesting because previous studies focused on the unmet medical needs of chronic drug users and unhealthy substance use (Al-Tayyib et al., 2015; Chitwood et al., 1999; Chitwood et al., 2001; Lasser et al., 2011; Metsch et al., 2002; Powelson et al., 2014; Robbins et al., 2010; Vu et al., 2015). The conclusions of these studies were that chronic drug users are a population in need of medical care. Thus, patterns of healthcare renunciation and access to medical care among experimenters and recreational drug users were unknown. It appears that all drug users, including those who are not heavy or
chronic users, are a vulnerable population at risk for forgone care. There was a gradient for tobacco and cannabis uses, more intensive use being associated with increased healthcare renunciation. The more participants smoked, the more likely they were to renounce care. Similarly, participants who used cannabis in the previous twelve months were more likely to renounce care than those who used cannabis in their lifetimes but not in the previous twelve months. There was no gradient for other illicit drug users, but the small prevalence of twelve-month users may have resulted in a lack of power (2.1%). Therefore, it seemed that whatever the drug use was, it was associated with increased healthcare renunciation. This information would be valuable for health policies, which should focus not only on high-risk people, i.e., heavy and chronic drug users but also on occasional users. Indeed, to our knowledge, no study reported that occasional or experimental drug users are at risk for health inequalities. Preventive programs should be developed to focus on this at-risk group and to prevent increases in people’s vulnerability because of experimental drug use. Additionally, further actions should include information on former use of substances, since people who used drugs in the past were also likelier to renounce care. It seems that there is no harmless substance use in term of health inequalities. In practice, clinicians and practitioners should be aware that all king of substance use, regardless the level of use, is likely to increase vulnerability.
6
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
Hazardous use of alcohol (binge drinking and being at risk for hazardous drinking) was associated with an increased healthcare renunciation, in line with the results of other substances. On the contrary, previous twelve-month use was not associated with increased forgone care in comparison with lifetime abstainers. Moderate alcohol use thus did not seem to be harmful in terms of healthcare renunciation. Participants who were alcohol users but not in the previous twelve months were on the contrary more likely to renounce care. It is well known that this subgroup of drinkers is heterogeneous and includes people who stopped drinking due to major health problems, such as alcohol use disorder or severe health problems (Dupuis et al., 2014). They are also more likely to use illicit drugs (Dupuis, Baggio, Accard, Mohler-Kuo, & Gmel, 2016). Our findings extended these results and showed that alcohol cessation was associated with other forms of vulnerability, namely healthcare renunciation. Former drinkers thus consistently appeared to be a high-risk group for health problems. Further studies should focus on former drinkers to achieve a better understanding of this heterogeneous subgroup. On the contrary, twelve-month gambling was not associated with healthcare renunciation when controlling for covariates, even for active gamblers who gambled at least once a week or spent 500s per year. However, the higher the score on the CGPI, which measures disordered gambling, the more likely participants were to forgo care. Therefore, addictive behaviors (i.e., without substance use) also seemed to be at risk for healthcare renunciation, even if gambling is not directly associated with adverse medical consequences such as substance use. Disordered gambling may be associated with financial problems and thus may result in higher rates of healthcare renunciation for economic reasons. This study suffered from some shortcomings. The first was that it was a cross-sectional study. Therefore, causal paths between substance use and healthcare renunciation cannot be tested. Further studies should investigate whether substance use results in increased forgone care or if these two variables are consequences of an important vulnerability. Second, other behavioral addictions, such as Internet addiction, gaming addiction, or workaholism, were not assessed. Assessing other addictions would be important to achieve a better understanding of the associations of healthcare renunciation with behavioral addictions. To conclude, it is already well known that socioeconomic factors influence healthcare renunciation and thus health inequalities. This study also showed that addictive behaviors, including substance use and gambling, were a part of the burden of vulnerability of people who forgo care. Therefore, drug use and gambling patterns should be focused on in the development of policies to reduce health inequalities. Conflict of interest The authors declare no conflict of interests. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or non-for-profit sectors. References Al-Tayyib, A. A., Thiede, H., Burt, R. D., & Koester, S. (2015). Unmet health care needs and hepatitis C infection among persons who inject drugs in Denver and Seattle, 2009. Prevention Science, 16(2), 330–340. http://dx.doi.org/10.1007/s11121-0140500-4.
Bodenmann, P., Favrat, B., Wolff, H., Guessous, I., Panese, F., Herzig, L., . . . Vaucher, P. (2014). Screening primary-care patients forgoing health care for economic reasons. PloS One, 9(4). http://dx.doi.org/10.1371/journal.pone.0094006. Chitwood, D. D., McBride, D. C., French, M. T., & Comerford, M. (1999). Health care need and utilization: A preliminary comparison of injection drug users, other illicit drug users, and nonusers. Substance Use & Misuse, 34(4–5), 727–746. http://dx.doi.org/10.3109/10826089909037240. Chitwood, D. D., Sanchez, J., Comerford, M., & McCoy, C. B. (2001). Primary preventive health care among injection drug users, other sustained drug users, and non-users. Substance Use & Misuse, 36(6–7), 807–824. Dawson, D. A. (2003). Methodological issues in measuring alcohol use. Alcohol Research & Health: The Journal of the National Institute on Alcohol Abuse and Alcoholism, 27(1), 18–29. Després, C., Dourgnon, P., Fantin, R., & Jusot, F. (2011). Giving up on health care for economic reasons: An econometric approach. Issues in Health Economics, 170, 1– 6. Dupuis, M., Baggio, S., Henchoz, Y., Deline, S., N’Goran, A., Studer, J., . . . Gmel, G. (2014). Risky single occasion drinking frequency and alcohol-related consequences: Can abstinence during early adulthood lead to alcohol problems? Swiss Medical Weekly, 144, w14017. http://dx.doi.org/10.4414/smw.2014.14017. Dupuis, M., Baggio, S., Accard, M. E., Mohler-Kuo, M., & Gmel, G. (2016). The association between alcohol abstinence, drinking or binge drinking and drug use: Is alcohol abstinence that safe? Drugs and Alcohol Today, 16(3), 212–221. http://dx.doi.org/10.1108/DAT-08-2015-0050. Ferris, J., Wynne, H. (2001). The Canadian Problem Gambling Index: final report. Retrieved from Ottawa. Guessous, I., Gaspoz, J. M., Theler, J. M., & Wolff, H. (2012). High prevalence of forgoing healthcare for economic reasons in Switzerland: A population-based study in a region with universal health insurance coverage. Preventive Medicine, 55(5), 521–527. http://dx.doi.org/10.1016/j.ypmed.2012.08.005. Guessous, I., Theler, J.-M., Izart, C. D., Stringhini, S., Bodenmann, P., Gaspoz, J.-M., & Wolff, H. (2014). Forgoing dental care for economic reasons in Switzerland: A six-year cross-sectional population-based study. BMC Oral Health, 14, 121. http://dx.doi.org/10.1186/1472-6831-14-121. Herring, R., Berridge, V., & Thom, B. (2008). Binge drinking: An exploration of a confused concept. Journal of Epidemiology & Community Health, 62(6), 476–479. http://dx.doi.org/10.1136/jech.2006.056721. Holtgraves, T. (2009). Evaluating the problem gambling severity index. Journal of Gambling Studies, 25(1), 105. http://dx.doi.org/10.1007/s10899-008-9107-7. Lasser, K. E., Kim, T. W., Alford, D. P., Cabral, H., Saitz, R., & Samet, J. H. (2011). Is unhealthy substance use associated with failure to receive cancer screening and flu vaccination? A retrospective cross-sectional study. BMJ Open, 1(1), e000046. http://dx.doi.org/10.1136/bmjopen-2010-000046. Legleye, S., Piontek, D., Kraus, L., Morand, E., & Falissard, B. (2013). A validation of the Cannabis Abuse Screening Test (CAST) using a latent class analysis of the DSMIV among adolescents. International Journal of Methods in Psychiatric Research, 22 (1), 16–26. http://dx.doi.org/10.1002/mpr.1378. Legleye, S., Guignard, R., Richard, J.-B., Ludwig, K., Pabst, A., & Beck, F. (2015). Properties of the Cannabis Abuse Screening Test (CAST) in the general population. International Journal of Methods in Psychiatric Research, 24(2), 170–183. http://dx.doi.org/10.1002/mpr.1465. Leplège, A., Ecosse, E., Verdier, A., & Perneger, T. V. (1998). The French SF-36 Health Survey: Translation, cultural adaptation and preliminary psychometric evaluation. Journal of Clinical Epidemiology, 51(11), 1013–1023. Metsch, L. R., Crandall, L., Wohler-Torres, B., Miles, C. C., Chitwood, D. D., & McCoy, C. B. (2002). Met and unmet need for dental services among active drug users in Miami, Florida. The Journal of Behavioral Health Services & Research, 29(2), 176– 188. Mouquet, M. C., & Villet, H. (2002). Les risques d’alcoolisation excessive chez les patients ayant recours aux soins un jour donné. Etudes et re´sultats, 192, 1–11. Myers, B. J. (2012). Primary health care for people who inject drugs in low and middle income countries. The International Journal on Drug Policy, 23(2), 105– 106. http://dx.doi.org/10.1016/j.drugpo.2011.09.014. Pampel, F. C., Krueger, P. M., & Denney, J. T. (2010). Socioeconomic disparities in health behaviors. Annual Review of Sociology, 36, 349–370. http://dx.doi.org/ 10.1146/annurev.soc.012809.102529. Pierre, A., & Jusot, F. (2017). The likely effects of employer-mandated complementary health insurance on health coverage in France. Health Policy, 121, 321–328. http://dx.doi.org/10.1016/j.healthpol.2016.12.004. Powelson, E., Lorvick, J., Lutnick, A., Wenger, L., Klausner, J., & Kral, A. H. (2014). Unmet healthcare need among women who use methamphetamine in San Francisco. Substance Use & Misuse, 49(3), 243–252. http://dx.doi.org/10.3109/ 10826084.2013.825919. Röttger, J., Blümel, M., Köppen, J., & Busse, R. (2016). Forgone care among chronically ill patients in Germany—Results from a cross-sectional survey with 15,565 individuals. Health Policy (Amsterdam, Netherlands), 120(2), 170–178. http://dx. doi.org/10.1016/j.healthpol.2016.01.004. Richard, J. B., Andler, R., Gautier, A., Guignard, R., Leon, C., & Beck, F. (2016). Effects of using an overlapping dual-frame design on estimates of health behaviors: A French general population telephone survey. Journal of Survey Statistics and Methodology, smw028, 1–21. Robbins, J. L., Wenger, L., Lorvick, J., Shiboski, C., & Kral, A. H. (2010). Health and oral health care needs and health care-seeking behavior among homeless injection drug users in San Francisco. Journal of Urban Health, 87(6), 920–930. http://dx. doi.org/10.1007/s11524-010-9498-5.
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7 Schoen, C., Osborn, R., Squires, D., Doty, M. M., Pierson, R., . . . Applebaum, S. (2010). How health insurance design affects access to care and costs, by income, in eleven countries. Health Affairs, 29(12), 2323–2334. http://dx.doi.org/10.1377/ hlthaff.2010.0862. Vu, F., Daeppen, J.-B., Hugli, O., Iglesias, K., Stucki, S., Paroz, S., . . . Bodenmann, P. (2015). Screening of mental health and substance users in frequent users of a
7
general Swiss emergency department. BMC Emergency Medicine, 15, 27. http:// dx.doi.org/10.1186/s12873-015-0053-2. Wolff, H., Gaspoz, J.-M., & Guessous, I. (2011). Health care renunciation for economic reasons in Switzerland. Swiss Medical Weekly, 141, w13165. http://dx.doi.org/ 10.4414/smw.2011.13165.
Contents lists available at ScienceDirect
International Journal of Drug Policy journal homepage: www.elsevier.com/locate/drugpo
Research paper
Addictive behaviors and healthcare renunciation for economic reasons in a French population-based sample Stéphanie Baggioa,f,* , Marc Dupuisb , Jean-Baptiste Richardc , François Beckd,e a
Life Course and Social Inequality Research Centre, University of Lausanne, Switzerland Institute of Psychology, University of Lausanne, Switzerland c Santé publique France, The French National Public Health Agency, France d Observatoire Français des Drogues et des Toxicomanies (OFDT), The French Monitoring Center for Drugs and Drug Addiction, France e ERES, Sorbonne Universités, UPMC Univ Paris 06, INSERM, Institut Pierre Louis d’Epidémiologie et de Santé Publique, UMRS, 1136 Paris, France f Division of Correctional Medicine, Geneva University Hospitals and University of Geneva, Geneva, Switzerland b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 February 2017 Received in revised form 22 April 2017 Accepted 10 July 2017 Available online xxx
Background: Healthcare renunciation for economic reasons is a major health concern, but it has been scarcely investigated among drug users, even if drug users constitute a vulnerable population in need of medical care. This study investigated associations of healthcare renunciation for economic reasons and addictive behaviors (alcohol, tobacco, cannabis, illicit drug use, and gambling) in a population-based sample of adults living in France, a country with universal health coverage. Methods: Data were collected using the 2014 Health Barometer, a French cross-sectional survey conducted among a random representative sample of the general population aged 18–64 (n = 12,852). Measures included healthcare renunciation, substance use (alcohol, tobacco, cannabis, and other illicit drugs) and gambling. Experimental/recreational and heavy/chronic use were assessed. Logistic regressions were used to test the relationship between healthcare renunciation and addictive behaviors, controlling for relevant covariates. Results: A total of 25% of the participants had renounced care at least once in the previous twelve months. Most variables of drug use were significantly associated with increased healthcare renunciation. This was the case for heavy/hazardous use and experimental/recreational use. Regular gambling was not associated with healthcare renunciation, but disordered gambling was. Conclusion: This study showed that addictive behaviors, including substance use and gambling, were part of the burden of vulnerability of people who forgo care. Therefore, drug use and gambling patterns should be a focus in the development of policies to reduce health inequalities, not only for heavy and chronic drug users. © 2017 Elsevier B.V. All rights reserved.
Keywords: Forgone care Gambling Population-based survey Substance use Health social inequalities
Introduction Healthcare renunciation for economic reasons, i.e., forgoing care due to financial reasons although clearly perceiving the need for it, is a major health concern in health inequality research. Vulnerable people are especially a focus because people in situations of deprivation and high-risk populations (e.g., those with chronic illnesses) are more likely to forgo care for economic reasons worldwide (Bodenmann et al., 2014; Guessous, Gaspoz, Theler, & Wolff, 2012; Pampel, Krueger, & Denney, 2010; Röttger,
* Corresponding author at: Institute for Social Sciences, University of Lausanne, Geopolis Building, CH-1015 Lausanne, Switzerland. E-mail addresses: [email protected] (S. Baggio), [email protected] (M. Dupuis), [email protected] (J.-B. Richard), [email protected] (F. Beck). http://dx.doi.org/10.1016/j.drugpo.2017.07.004 0955-3959/© 2017 Elsevier B.V. All rights reserved.
Blümel, Köppen, & Busse, 2016; Wolff, Gaspoz, & Guessous, 2011). Previous studies have mainly focused on the relationships of forgone care with socioeconomic status, demographics, and perceived health status. Studies found consistent associations of healthcare renunciation for economic reasons of having low income or job position, being older, being female, having dependent children, being divorced or single, and having bad health outcomes in different countries (Bodenmann et al., 2014; Guessous et al., 2012; Pampel et al., 2010; Röttger et al., 2016; Wolff et al., 2011). Another part of the literature on the utilization of health care services highlighted that drug users constitute a vulnerable population in need of medical care. In several countries, drug users have been described as having an important lack of access to primary care (Myers, 2012) and are considered medically underserved (Metsch et al., 2002), although addiction is associated
2
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
with adverse medical consequences. Studies have mostly focused on high-risk drug users’ unmet healthcare needs, such as injecting drug users (Al-Tayyib, Thiede, Burt, & Koester, 2015; Chitwood, McBride, French, & Comerford, 1999; Metsch et al., 2002; Robbins, Wenger, Lorvick, Shiboski, & Kral, 2010) and methamphetamine users (Powelson et al., 2014). This population has a large burden of unmet health needs and is more likely to visit hospital emergency departments (Vu et al., 2015). However, information on associations of occasional or recreational drug use and other addictive behaviors such as gambling with unmet health care in the general population is lacking. Some studies reported that, in the general population, unhealthy substance use was a risk factor for not receiving preventive healthcare, such as breast-cancer screening, flu vaccination, and control visits to general practitioners (Chitwood, Sanchez, Comerford, & McCoy, 2001; Lasser et al., 2011), but investigations on the relationship between forgone care and experimental or recreational drug use are needed. Overall, if healthcare renunciation is a topic of growing interest, to our knowledge, no study has focused specifically on the association between addictive behaviors and healthcare renunciation in the general population. Because drug users are likely to have unmet healthcare needs, they are susceptible to be a high-risk population for healthcare renunciation. For example, Guessous et al. (2012) reported that current smokers were more likely to renounce care than non-smokers, and Guessous et al. (2014) reported that current smokers were more likely to renounce dental care than nonsmokers. However, these studies did not report other substance use, and no information on the addictive level of tobacco use was available. This study aimed to fill this gap and investigated associations of healthcare renunciation for economic reasons and addictive behaviors (alcohol, tobacco, cannabis, other illicit drug use, and gambling) in a population-based sample of French adults. France has a universal health system, with a national public health insurance based on income. It includes significant cost sharing (approximately 30% of health care costs). Cost sharing is eliminated for individuals with several specified chronic conditions and lowincome individuals (in 2017, the upper limit for cost sharing elimination is 727s [$778] for one person, 1090s [$1167] for two persons). It is also lowered or eliminated for highly effective prescription drugs (Schoen et al., 2010). However, out-of-pocket expenditures may affect even low-income individuals (e.g., one single person having more than 727s per month). Most French citizens buy private complementary insurance that covers cost sharing (Pierre & Jusot, 2017). Data on healthcare renunciation are scarce in France. Methods Participants and procedure Data were collected using the 2014 Health Barometer, a French cross-sectional survey conducted among a random representative sample of the French population aged 15–75 years between December 2013 and May 2014. It is conducted using computerassisted telephone interviews. Telephone numbers of households (landlines and cell phones) were randomly generated, and individuals in households were randomly selected among eligible household members: having a member between 15 and 75 years old, speaking French, and living within the household during the survey. When the selected individual was unavailable, an appointment was made. Households or individuals were considered unreachable after 40 telephone calls. The response rate was 61% for landlines and 52% for cell phones. The sample included 15,635 participants, with n = 7577 for landlines and n = 8058 for cell phones. This study focused on an adult population that was 18–
64 years old (n = 13,039). We excluded participants older than 64 (65–75 years old) because they did not complete the questions related to illicit drug use. Missing values on the variables included in the study were listwise deleted (187 participants had missing values), leaving a final sample of n = 12,852. Weights were constructed in two steps: the design weight computation (which reflects the differential probabilities of selection, depending on several factors, including probability of selection of the telephone number, number of eligible individuals living in the households, number of landline and cellular telephones owned by the respondent), followed by an overall post-stratification to correct under-coverage. The questionnaire included questions on demographics, health behaviors, health status, and social determinants. The questionnaire, method, and evaluation of the method are available elsewhere (Richard et al., 2016). Measures Healthcare renunciation Participants answered whether they had renounced care for financial reasons using the following dimensions for the previous 12 months: (1) dental care, (2) eye care, (3) medical appointment, and (4) other health care. Healthcare renunciation was coded 1 if participants renounced at least to one situation in which healthcare was needed and 0 otherwise. We assessed lifetime and 12-month prevalence of substance use and gambling patterns. A 12-month period has been recommended to assess substance use behavior and problems (Dawson, 2003). Alcohol use Alcohol use was assessed using three variables: (1) alcohol use with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use; (2) binge drinking, defined as having drunk at least six drinks in a single session in the previous month (yes/no) (Herring, Berridge, & Thom, 2008), and (3) hazardous drinking (at risk/not at risk), defined usually as drinking more than 21 drinks a week for a man and 14 for a woman or drinking six drinks or more on a single occasion weekly (Mouquet & Villet, 2002). Tobacco use Tobacco use was assessed using a variable related to smoking status with five categories: no lifetime tobacco use, lifetime smoker but currently non-smoker, occasional smoker (smoking but not daily), daily smoker (with less than ten cigarettes per day), and heavy smoker (daily smoker with ten cigarettes or more per day). We classified participants who had smoked only once or twice in their lives as non-smokers. Cannabis use Cannabis use was assessed with two variables: (1) cannabis use with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use and (2) score on the cannabis abuse screening test (CAST, Legleye et al., 2015), ranging from 0 to 24, with a larger score indicating cannabis-use disorder. Previous studies reported good psychometric properties of the CAST (Legleye et al., 2013; Legleye, Piontek, Kraus, Morand, & Falissard, 2013). Other illicit drug use Participants indicated whether they used any illicit drug use other than cannabis (one question for each of the following illicit drugs: magic mushrooms, poppers, other inhalants, ecstasy,
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
3
Table 1 Descriptive statistics and bivariate associations of healthcare renunciation, addictive behaviors, and covariates. Variables
Descriptive
Healthcare renunciation No 75.0% (n = 9636)
Yes 25.0% (n = 3219)
p-value
Effect size
Age1
42.2 (13.1)
42.3
41.7
.081
–
Gender2 Male Female
47.0 53.0
79.9a 70.6b
20.1a 29.4b
<.001
.107
Living area2 Rural Urban
26.0 74.0
78.6a 73.7b
21.4a 26.3b
<.001
.049
Socioeconomic category2 Farm worker Craftsman/retailer/business owner Manual worker Employee Intermediate occupation Executive/senior-level occupation Do not know/do not answer
1.6 5.5 19.3 27.1 25.6 19.1 1.8
85.2ab 78.5ab 72.2c 68.4d 76.6a 83.1b 74.9acd
14.8ab 21.5a 27.8c 31.6d 23.4a 16.9b 25.1acd
<.001
.124
Income2 First quintile Second quintile Third quintile Forth quintile Fifth quintile Do not known/do not want to answer
14.8 17.7 17.7 22.0 23.2 4.6
60.9a 65.7b 72.6c 79.2d 88.0e 79.3d
39.1a 34.3b 27.4c 20.8d 12.0e 20.7d
<.001
.218
Subjective health status2 Good/average Bad Mental health status (0–100)1
94.6 5.4 69.5 (17.3)
76.2a 53.3b 71.7
23.8a 46.7b 63.0
<.001
.119
<.001
.047
Alcohol use2 No lifetime use Lifetime use but not current user 12-month use
3.3 7.4 89.3
64.8a 65.6a 76.1b
35.2a 34.4a 23.9b
<.001
.077
Binge drinkingb (among 12-month alcohol user, n = 11,481) No Yes (monthly or more)
80.2 19.8
76.5a 74.3b
23.5a 25.7b
.104
–
Hazardous drinking2 (among 12-month alcohol user, n = 11,481) Not at risk At risk
52.1 47.9
75.8a 76.5a
24.2a 23.5a
.417
–
Tobacco use2 No lifetime use Lifetime use but not current user Occasional smoker Daily smoker Heavy smoker (>10 cigarettes/day)
<.001
.120
65.4 9.4 6.5 18.7
78.2a 76.4ab 71.2b 65.0c
21.8a 23.6ab 28.8b 35.0c
Cannabis use2 No lifetime use Lifetime use but not current user 12-month use CAST (0–12)3 (among 12-month cannabis user, n = 1219)
57.5 32.6 9.8 2.9 (3.7)
77.9a 72.7b 65.2c 2.5
22.1a 27.3b 34.8c 3.7
<.001
.093
<.001
–
<.001
.069
.834
–
<.001
–
Illicit drug use2 No lifetime use Lifetime use but not current user 12-month use
86.9 11.0 2.1
76.1a 68.0b 64.3b
23.9a 32.0b 35.7b
Gambling2 No 12-month use 12-month use Active use (once a week or 500s/year) CPGI (0–24)3 (among 12-month gamblers, n = 7667)
40.4 41.9 17.7 0.3 (1.2)
74.8a 75.2a 74.6a 0.26
25.2a 24.8a 25.4a 0.42
CAST: Cannabis Abuse Screening Test, CPGI: Canadian Problem Gambling Index. a,b,c,d,e A same subscript letter within a column denotes that proportions did not differ; two different subscript letters denote that proportions differed at the .05 level. A Bonferroni–Holms correction was applied to keep a 5% error rate. Significant p-values were adjusted. 1 Means and standard deviations are given. t-tests for bivariate associations were computed, with partial eta squares. 2 Percentages are given. Chi-square tests with Fisher exact tests to compare columns proportions and Cramer V are given for bivariate associations. Line percentages are given for bivariate associations. No effect size was provided for non-significant associations. 3 Means and standard deviations are given. Negative binomial regressions were used for bivariate associations. No effect size is available.
4
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
amphetamines, methamphetamines, LSD, crack, coke, heroine, and GHB/GHL) with three categories: no lifetime use, lifetime use but not in the previous twelve months, and previous twelve-months use. Gambling Gambling was assessed using two variables: (1) gambling with three categories: no previous twelve-month use, previous twelvemonths use, and active gambling, defined as having gambled at least once a week or having spent 500s on gambling the previous year and (2) score on the Canadian Problem Gambling Index (CPGI, Ferris & Wynne, 2001), ranging from 0 to 24, with a larger score indicating disordered gambling. The CGPI has been described as a reliable instrument for non-clinical contexts (Holtgraves, 2009). Covariates Participants provided information on their age, sex, living area (urban versus rural), socioeconomic category (farm worker, manual worker, employee, intermediate occupation, executive/ senior-level professional occupation, craftsman/retailer/business owner, and do not know/do not want to answer), and household income per consumption unit. Consumption units were used to compare households of different sizes and composition by assigning a coefficient to each member of the household. We used the OECD scale to calculate the consumption units (one unit for the first adult, 0.5 unit for individuals aged 14 and older, and 0.3 for children under 14 years). Income was recoded in quintiles, with an additional category do not know/do not want to answer. Participants also evaluated their health (subjective health status), which was coded as good/average or bad. Mental health status was assessed for the previous four weeks with five questions (nervousness, discouragement, calm/relaxation, sadness, and happiness) assessing the MH-5, which is a subscale of the mental health subscales of the SF-36 with good psychometric properties (Leplège, Ecosse, Verdier, & Perneger, 1998) and computed to create a mean score ranging from 0 (bad mental health) to 100 (good mental health). Statistical analyses Descriptive statistics for healthcare renunciation, addictive behaviors, and covariates were computed using percentages and means according to the variables’ distribution. Then, bivariate associations between healthcare renunciation and all other variables were performed using t tests (continuous variables), negative binomial regressions (continuous skewed variables), and chi-square tests with pairwise comparisons using Fisher’s exact tests (categorical variables). Effect sizes were computed using partial eta squares for continuous variables and Cramer’s V for categorical variables. Then, multivariate associations of healthcare renunciation and addictive behaviors were computed, controlling for covariates. We computed logistic regressions with healthcare renunciation as the dependent variable and each addictive behavior variable as the independent variable. We reported individual coefficients and statistics on the overall goodness-offit (difference between the deviance of the model without the variable of interest and with the variable of interest). We did not perform a model including all addictive behavior-related variables together to avoid collinearity problems. A Bonferroni-Holm correction to maintain a 5% error rate was used for bivariate and multivariate associations. All analyses were performed using SPSS 24. Results Descriptive statistics and bivariate associations are reported in Table 1. A total of 25.0% of the participants had renounced care for
financial reasons at least once in the previous twelve months. More precisely, 16.4% renounced dental care, 11.1% renounced eye care, 7.1% renounced a medical appointment, and 9.5% renounced other another form of health care, with an average of 1.76 0.97 dimensions of forgoing care among those who renounced care. Demographics were associated with healthcare renunciation (first panel of Table 1): being a woman (p < .001), living in an urban area (p < .001), having a low socioeconomic status (p < .001), and having a low income (p < .001). Subjective health and mental health status were also associated with healthcare renunciation: participants who renounced care had a lower (mental) health score (p < .001). The highest effect size was for income (.218). Age was not significantly associated with healthcare renunciation (p = .081). Addictive behaviors were also associated with healthcare renunciation (second panel of Table 1). Tobacco, cannabis, and other illicit drug use were associated with increased healthcare renunciation (p < .001). Alcohol (binge drinking and hazardous drinking) was not associated with healthcare renunciation, except previous 12-month alcohol use, which was associated with decreased healthcare renunciation in comparison with lifetime and previous twelve months abstainers (p < .001). Twelve-month gambling was not associated with healthcare renunciation, but CPGI score was, with participants who renounced healthcare having a higher score of disordered gambling (p < .001). The effect sizes were small, ranging from .069 to .120 for significant associations. In multivariate associations, reported in Table 2, the results remained mostly unchanged, with substance users (tobacco, cannabis, and other illicit drugs) being more likely to renounce care, with a gradient for tobacco and cannabis. The results were also similar to bivariate associations for gambling, with a marginal increased risk of healthcare renunciation when participants had a higher CGPI score (p = .054) but not twelve-month gambling. Binge drinkers were significantly more likely to renounce care, contrary to bivariate associations that displayed non-significant associations. Twelve-month alcohol users were less likely to renounce care, but they were not significantly different from lifetime abstainers. Former drinkers, who were lifetime drinkers but had not drunk in the previous twelve months were more likely to renounce care than current alcohol users. For all models except twelve-month gambling, the model including the variable of interest was better than the model that did not include it (i.e., lower deviance). Discussion This study investigated associations of healthcare renunciation for economic reasons and addictive behaviors among a populationbased sample of people living in France with Universal Health Coverage. First, a total of 25.0% of the participants renounced care at least once in the previous twelve months, regardless of the kind of renunciation (dental care, eye care, medical appointments, and other kinds of health care). This prevalence rate is higher than those previously reported in France (15.4% among the adult population in 2008; Després, Dourgnon, Fantin, & Jusot, 2011). Forgoing care for financial reasons has increased between 2002 and 2008 (Després et al., 2011), and it seems that it has increased since then. This possible increase is worrying information. Additionally, healthcare renunciation remained a factor of social inequalities because people with low income, low socioeconomic status, and bad mental health outcomes were more likely to renounce care, as was the case in the previous decade (Després et al., 2011).
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
5
Table 2 Logistic regressions of healthcare renunciation predicted by addictive behaviors, controlling for covariates. Variables (IV)
Healthcare renunciation (DV)
p-value3
Difference in deviance4
Alcohol use2 No lifetime use Lifetime use but not current user 12-month use
0.25ab 0.27a 0.22b
.018
13.7
Binge drinking2 (among 12-month alcohol user, n = 11,481) No Yes (monthly or more)
0.22a 0.25b
.020
13.0
Hazardous drinking2 (among 12-month alcohol user, n = 11,481) Not at risk At risk
0.22a 0.24a
.063
26.9
Tobacco use2 No lifetime use Lifetime use but not current user Occasional smoker Daily smoker Heavy smoker (>10 cigarettes/day)
0.18a 0.21b 0.23bc 0.26c 0.30d
<.001
117.0
Cannabis use2 No lifetime use Lifetime use but not current user 12-month use CAST (0–12)1 (among 12-month cannabis user, n = 1219)
0.19a 0.27b 0.35c 0.49
<.001
162.6
.024
70.3
Illicit drug use No lifetime use Lifetime use but not current user 12-month use
0.22a 0.32b 0.34b
<.001
79.6
Gambling2 No 12-month use 12-month use Active use (once a week or 500 s/year) CPGI (0–24)1 (among 12-month gamblers, n = 7667)
0.23a 0.23a 0.23a 0.047
.797
0.3
.054
5.5
2
CAST: Cannabis Abuse Screening Test, CPGI: Canadian Problem Gambling Index. Dependent variable (DV): healthcare renunciation with reference category ‘no healthcare renunciation’, independent variable (IV): addictive behaviors. Separate logistic regressions were run to avoid multicollinearity, with: 1 unstandardized b for continuous IV (CAST and CPGI), and 2 adjusted means and pairwise comparisons for categorical IV, 3 p-values of Wald Chi-squares are given. 4 Differences in deviance (-2 loglikelihood) between the model without the IV of interest (but with control variables) and with the IV of interest are reported. a,b,c,d A same subscript letter within a column denotes that proportions did not differ; two different subscript letters denote that proportions differed at the .05 level. Models controlled for age, gender, socioeconomic status, income, health status, and mental health status. A Bonferroni-Holms correction was applied to keep a 5% error rate.
Addictive behaviors added to the burden of these already vulnerable individuals. Indeed, substance and hazardous substance use were associated with healthcare renunciation. Controlling for important covariates (demographics and mental health status), almost all tested variables were significantly associated with healthcare renunciation. Heavy and hazardous substance use, defined with heavy smoking (ten cigarettes per day or more), binge drinking (six or more drinks on a single occasion per month), being at risk for hazardous drinking, and having a high score on the cannabis-use disorder scale were associated with increased healthcare renunciation. Lifetime use of tobacco, cannabis, and other illicit drugs were also associated with increased forgone care. Therefore, experimental and recreational use seemed associated with healthcare renunciation and not only intensive and hazardous use, as previously shown. This result was interesting because previous studies focused on the unmet medical needs of chronic drug users and unhealthy substance use (Al-Tayyib et al., 2015; Chitwood et al., 1999; Chitwood et al., 2001; Lasser et al., 2011; Metsch et al., 2002; Powelson et al., 2014; Robbins et al., 2010; Vu et al., 2015). The conclusions of these studies were that chronic drug users are a population in need of medical care. Thus, patterns of healthcare renunciation and access to medical care among experimenters and recreational drug users were unknown. It appears that all drug users, including those who are not heavy or
chronic users, are a vulnerable population at risk for forgone care. There was a gradient for tobacco and cannabis uses, more intensive use being associated with increased healthcare renunciation. The more participants smoked, the more likely they were to renounce care. Similarly, participants who used cannabis in the previous twelve months were more likely to renounce care than those who used cannabis in their lifetimes but not in the previous twelve months. There was no gradient for other illicit drug users, but the small prevalence of twelve-month users may have resulted in a lack of power (2.1%). Therefore, it seemed that whatever the drug use was, it was associated with increased healthcare renunciation. This information would be valuable for health policies, which should focus not only on high-risk people, i.e., heavy and chronic drug users but also on occasional users. Indeed, to our knowledge, no study reported that occasional or experimental drug users are at risk for health inequalities. Preventive programs should be developed to focus on this at-risk group and to prevent increases in people’s vulnerability because of experimental drug use. Additionally, further actions should include information on former use of substances, since people who used drugs in the past were also likelier to renounce care. It seems that there is no harmless substance use in term of health inequalities. In practice, clinicians and practitioners should be aware that all king of substance use, regardless the level of use, is likely to increase vulnerability.
6
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7
Hazardous use of alcohol (binge drinking and being at risk for hazardous drinking) was associated with an increased healthcare renunciation, in line with the results of other substances. On the contrary, previous twelve-month use was not associated with increased forgone care in comparison with lifetime abstainers. Moderate alcohol use thus did not seem to be harmful in terms of healthcare renunciation. Participants who were alcohol users but not in the previous twelve months were on the contrary more likely to renounce care. It is well known that this subgroup of drinkers is heterogeneous and includes people who stopped drinking due to major health problems, such as alcohol use disorder or severe health problems (Dupuis et al., 2014). They are also more likely to use illicit drugs (Dupuis, Baggio, Accard, Mohler-Kuo, & Gmel, 2016). Our findings extended these results and showed that alcohol cessation was associated with other forms of vulnerability, namely healthcare renunciation. Former drinkers thus consistently appeared to be a high-risk group for health problems. Further studies should focus on former drinkers to achieve a better understanding of this heterogeneous subgroup. On the contrary, twelve-month gambling was not associated with healthcare renunciation when controlling for covariates, even for active gamblers who gambled at least once a week or spent 500s per year. However, the higher the score on the CGPI, which measures disordered gambling, the more likely participants were to forgo care. Therefore, addictive behaviors (i.e., without substance use) also seemed to be at risk for healthcare renunciation, even if gambling is not directly associated with adverse medical consequences such as substance use. Disordered gambling may be associated with financial problems and thus may result in higher rates of healthcare renunciation for economic reasons. This study suffered from some shortcomings. The first was that it was a cross-sectional study. Therefore, causal paths between substance use and healthcare renunciation cannot be tested. Further studies should investigate whether substance use results in increased forgone care or if these two variables are consequences of an important vulnerability. Second, other behavioral addictions, such as Internet addiction, gaming addiction, or workaholism, were not assessed. Assessing other addictions would be important to achieve a better understanding of the associations of healthcare renunciation with behavioral addictions. To conclude, it is already well known that socioeconomic factors influence healthcare renunciation and thus health inequalities. This study also showed that addictive behaviors, including substance use and gambling, were a part of the burden of vulnerability of people who forgo care. Therefore, drug use and gambling patterns should be focused on in the development of policies to reduce health inequalities. Conflict of interest The authors declare no conflict of interests. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or non-for-profit sectors. References Al-Tayyib, A. A., Thiede, H., Burt, R. D., & Koester, S. (2015). Unmet health care needs and hepatitis C infection among persons who inject drugs in Denver and Seattle, 2009. Prevention Science, 16(2), 330–340. http://dx.doi.org/10.1007/s11121-0140500-4.
Bodenmann, P., Favrat, B., Wolff, H., Guessous, I., Panese, F., Herzig, L., . . . Vaucher, P. (2014). Screening primary-care patients forgoing health care for economic reasons. PloS One, 9(4). http://dx.doi.org/10.1371/journal.pone.0094006. Chitwood, D. D., McBride, D. C., French, M. T., & Comerford, M. (1999). Health care need and utilization: A preliminary comparison of injection drug users, other illicit drug users, and nonusers. Substance Use & Misuse, 34(4–5), 727–746. http://dx.doi.org/10.3109/10826089909037240. Chitwood, D. D., Sanchez, J., Comerford, M., & McCoy, C. B. (2001). Primary preventive health care among injection drug users, other sustained drug users, and non-users. Substance Use & Misuse, 36(6–7), 807–824. Dawson, D. A. (2003). Methodological issues in measuring alcohol use. Alcohol Research & Health: The Journal of the National Institute on Alcohol Abuse and Alcoholism, 27(1), 18–29. Després, C., Dourgnon, P., Fantin, R., & Jusot, F. (2011). Giving up on health care for economic reasons: An econometric approach. Issues in Health Economics, 170, 1– 6. Dupuis, M., Baggio, S., Henchoz, Y., Deline, S., N’Goran, A., Studer, J., . . . Gmel, G. (2014). Risky single occasion drinking frequency and alcohol-related consequences: Can abstinence during early adulthood lead to alcohol problems? Swiss Medical Weekly, 144, w14017. http://dx.doi.org/10.4414/smw.2014.14017. Dupuis, M., Baggio, S., Accard, M. E., Mohler-Kuo, M., & Gmel, G. (2016). The association between alcohol abstinence, drinking or binge drinking and drug use: Is alcohol abstinence that safe? Drugs and Alcohol Today, 16(3), 212–221. http://dx.doi.org/10.1108/DAT-08-2015-0050. Ferris, J., Wynne, H. (2001). The Canadian Problem Gambling Index: final report. Retrieved from Ottawa. Guessous, I., Gaspoz, J. M., Theler, J. M., & Wolff, H. (2012). High prevalence of forgoing healthcare for economic reasons in Switzerland: A population-based study in a region with universal health insurance coverage. Preventive Medicine, 55(5), 521–527. http://dx.doi.org/10.1016/j.ypmed.2012.08.005. Guessous, I., Theler, J.-M., Izart, C. D., Stringhini, S., Bodenmann, P., Gaspoz, J.-M., & Wolff, H. (2014). Forgoing dental care for economic reasons in Switzerland: A six-year cross-sectional population-based study. BMC Oral Health, 14, 121. http://dx.doi.org/10.1186/1472-6831-14-121. Herring, R., Berridge, V., & Thom, B. (2008). Binge drinking: An exploration of a confused concept. Journal of Epidemiology & Community Health, 62(6), 476–479. http://dx.doi.org/10.1136/jech.2006.056721. Holtgraves, T. (2009). Evaluating the problem gambling severity index. Journal of Gambling Studies, 25(1), 105. http://dx.doi.org/10.1007/s10899-008-9107-7. Lasser, K. E., Kim, T. W., Alford, D. P., Cabral, H., Saitz, R., & Samet, J. H. (2011). Is unhealthy substance use associated with failure to receive cancer screening and flu vaccination? A retrospective cross-sectional study. BMJ Open, 1(1), e000046. http://dx.doi.org/10.1136/bmjopen-2010-000046. Legleye, S., Piontek, D., Kraus, L., Morand, E., & Falissard, B. (2013). A validation of the Cannabis Abuse Screening Test (CAST) using a latent class analysis of the DSMIV among adolescents. International Journal of Methods in Psychiatric Research, 22 (1), 16–26. http://dx.doi.org/10.1002/mpr.1378. Legleye, S., Guignard, R., Richard, J.-B., Ludwig, K., Pabst, A., & Beck, F. (2015). Properties of the Cannabis Abuse Screening Test (CAST) in the general population. International Journal of Methods in Psychiatric Research, 24(2), 170–183. http://dx.doi.org/10.1002/mpr.1465. Leplège, A., Ecosse, E., Verdier, A., & Perneger, T. V. (1998). The French SF-36 Health Survey: Translation, cultural adaptation and preliminary psychometric evaluation. Journal of Clinical Epidemiology, 51(11), 1013–1023. Metsch, L. R., Crandall, L., Wohler-Torres, B., Miles, C. C., Chitwood, D. D., & McCoy, C. B. (2002). Met and unmet need for dental services among active drug users in Miami, Florida. The Journal of Behavioral Health Services & Research, 29(2), 176– 188. Mouquet, M. C., & Villet, H. (2002). Les risques d’alcoolisation excessive chez les patients ayant recours aux soins un jour donné. Etudes et re´sultats, 192, 1–11. Myers, B. J. (2012). Primary health care for people who inject drugs in low and middle income countries. The International Journal on Drug Policy, 23(2), 105– 106. http://dx.doi.org/10.1016/j.drugpo.2011.09.014. Pampel, F. C., Krueger, P. M., & Denney, J. T. (2010). Socioeconomic disparities in health behaviors. Annual Review of Sociology, 36, 349–370. http://dx.doi.org/ 10.1146/annurev.soc.012809.102529. Pierre, A., & Jusot, F. (2017). The likely effects of employer-mandated complementary health insurance on health coverage in France. Health Policy, 121, 321–328. http://dx.doi.org/10.1016/j.healthpol.2016.12.004. Powelson, E., Lorvick, J., Lutnick, A., Wenger, L., Klausner, J., & Kral, A. H. (2014). Unmet healthcare need among women who use methamphetamine in San Francisco. Substance Use & Misuse, 49(3), 243–252. http://dx.doi.org/10.3109/ 10826084.2013.825919. Röttger, J., Blümel, M., Köppen, J., & Busse, R. (2016). Forgone care among chronically ill patients in Germany—Results from a cross-sectional survey with 15,565 individuals. Health Policy (Amsterdam, Netherlands), 120(2), 170–178. http://dx. doi.org/10.1016/j.healthpol.2016.01.004. Richard, J. B., Andler, R., Gautier, A., Guignard, R., Leon, C., & Beck, F. (2016). Effects of using an overlapping dual-frame design on estimates of health behaviors: A French general population telephone survey. Journal of Survey Statistics and Methodology, smw028, 1–21. Robbins, J. L., Wenger, L., Lorvick, J., Shiboski, C., & Kral, A. H. (2010). Health and oral health care needs and health care-seeking behavior among homeless injection drug users in San Francisco. Journal of Urban Health, 87(6), 920–930. http://dx. doi.org/10.1007/s11524-010-9498-5.
S. Baggio et al. / International Journal of Drug Policy 49 (2017) 1–7 Schoen, C., Osborn, R., Squires, D., Doty, M. M., Pierson, R., . . . Applebaum, S. (2010). How health insurance design affects access to care and costs, by income, in eleven countries. Health Affairs, 29(12), 2323–2334. http://dx.doi.org/10.1377/ hlthaff.2010.0862. Vu, F., Daeppen, J.-B., Hugli, O., Iglesias, K., Stucki, S., Paroz, S., . . . Bodenmann, P. (2015). Screening of mental health and substance users in frequent users of a
7
general Swiss emergency department. BMC Emergency Medicine, 15, 27. http:// dx.doi.org/10.1186/s12873-015-0053-2. Wolff, H., Gaspoz, J.-M., & Guessous, I. (2011). Health care renunciation for economic reasons in Switzerland. Swiss Medical Weekly, 141, w13165. http://dx.doi.org/ 10.4414/smw.2011.13165.