Diabetes causal attributions: Pathways to stigma and health

Abstract

This study investigates the relationships among causal attributions, internalized stigma, and self-blame, along with downstream health and life satisfaction consequences for individuals with type 1 and type 2 diabetes. Data were analyzed from the Diabetes, Identity, Attributions, and Health study. Participants diagnosed with either type 1 or type 2 diabetes (N=363) were included in the analysis. Results indicated that the relationship between causal attributions and stigmatization was moderated by diabetes type. Path analyses, one for each diabetes type, revealed overall patterns linking causal attributions to internalized stigma and to self-blame, which were linked to ratings of reduced self-care, increased symptoms, and reduced life satisfaction. However, the specific paths diverged by diabetes type in important ways. Whereas higher genetic causal attributions were associated with more self-blame and stigmatization for type 1 diabetes, these attributions were associated with less self-blame and stigmatization for type 2 diabetes. The current work demonstrates the importance of causal attributions to overall health and illustrates how even in conditions with genetic attributions that are similar in magnitude, affected individuals may attach very different meaning to those attributions.

Introduction

Type 1 (T1D) and type 2 diabetes (T2D) are separate diseases that vary by etiology, prevalence, age of onset, risk factors, symptomatology, prevention, and treatment. A primary difference between diabetes types is their pattern of causal factors. Both are multifactorial in nature and are influenced by genetics, but only T2D is additionally associated with well-established behavioral risk factors. We propose that patients’ causal attributions for their disease have the potential to initiate beliefs and mindsets about the self that may ultimately influence health and life satisfaction outcomes. The current study therefore investigates how causal attributions, particularly to genetic and behavioral factors for T1D and T2D, are related to self-blame, internalized stigma, health, and life satisfaction outcomes among affected individuals.

Causal Foundations of Diabetes

Generally, the causal foundations of T1D are less clearly delineated than those of T2D. Estimated heritability for T1D ranges widely, with twin studies suggesting heritability of <50%, while research on T2D suggests a stronger heritable component of approximately 70% (Dean & McEntyre, 2004). Behavioral and environmental influences are more clearly divergent, particularly because those for T1D are not clearly defined. Posited environmental risk factors for T1D include cold weather environment and viral infections (National Institute of Diabetes and Digestive Kidney Diseases, 2016). Environmental causes of T2D include physical inactivity and/or a diet of processed foods/meats, saturated fats, and sugary beverages (Schnurr et al., 2020).

Existing alongside this objective scientific knowledge of causal factors in T1D and T2D are affected individuals’ perceptions of what causes their disease. Although causal attributions for these diseases are multifaceted (Rose, et al., 2019; French, Senior, Weinman, & Marteau, 2001), much of the research in this area focuses on contrasts between genetic (or biological) causal attributions and behavioral (or lifestyle) attributions. Perceptions about the causal foundations of diabetes vary considerably among affected individuals. Individuals with T1D tend to place a stronger emphasis on germs, viruses, genetics, and chance compared to behavioral factors in causing their disease (Rose, Costabile, Cohen, Boland, & Persky 2019). Qualitative work suggests that individuals with T1D specifically point to the lack of behavioral cause in their disease, which is often contrasted with T2D (Balfe et al., 2013). Meanwhile, individuals with T2D typically attribute both genes and behavior relatively equally in causing their own condition (Rose et al., 2019; Claassen et al., 2011). Additional factors such as body weight, the physical environment, and the family home environment are also frequently attributed as causes of T2D by affected individuals (Rose et al., 2019). As such, personal behavior is the primary factor on which T1D and T2D attributions tend to differ.

Influence of Causal Attributions on Blame and Stigma

One reason why these causal attributions are so important is due to their relationship with blame and stigmatization. In a general sense, theory has suggested that when a disease is thought to be caused by controllable factors (e.g., lifestyle behaviors), affected individuals are often held responsible and are stigmatized for their condition (Weiner, 1993; Weiner, Perry, & Magnusson, 1988). Conversely, evoking low control causes for a disease (e.g., genetics) reduces blame and stigma (Crandall, 1994; Weiner et al., 1988). In the context of obesity, when genetics is introduced as a causal factor for weight, stigmatization and blame toward affected individuals for causing their condition is reduced because genetics is perceived as uncontrollable (Hilbert 2016; Pearl & Lebowitz, 2014; Persky & Eccleston, 2011). Similar patterns of genetic attributions reducing blame and stigma have also been seen in addiction and mental illness (Lebowitz & Appelbaum, 2017; Kvaale, Gottdiener, & Haslam, 2013; Bennett, Thirlaway, & Murray, 2008; Phelan, Cruz-Rojas, & Reiff, 2002).

Endorsement of genetic causes does not necessarily imply a rejection of behavioral causes; the two often coexist (Sanderson, Waller, Humphries, & Wardle, 2011). When individuals hold behavioral causal attributions for a condition, this typically implies the condition is controllable. In the context of obesity, high-control attributions such as personal responsibility for weight have been associated with increased stigmatization, fat phobia, and weight bias internalization among affected individuals (Khan, Tarrant, Weston, Shah & Farrow, 2017; Pearl & Lebowitz, 2014). This notion of personal fault is a primary reason why stigmatized individuals often accept and endorse the stereotypes, prejudice, and discrimination leveled at their own stigmatized group (Alimoradi et al., 2020). This stigma can become internalized and manifest through self-stigmatization, which is especially pernicious (Corrigan, Watson & Barr, 2006).

In addition to controllability, causal factors also vary in terms of their perceived stability (Weiner et al., 1988). Compared to behavioral causal factors, genetic causal factors for health conditions may lead to judgments that the condition is relatively immutable (Dar-Nimrod & Heine, 2011). Genetic attributions for illness can thus encourage essentialist thinking because affected individuals are perceived to be fundamentally different from ‘healthy’ people (Haslam, 2011). Those affected by a condition considered to be genetic in origin may consequently be stigmatized due to beliefs that negative illness characteristics are a core and unchangeable part of the individual. Indeed, biological attributions for mental illness have been associated with increased desire for social distance from affected individuals and with beliefs that affected individuals may be inherently dangerous (Kvaale et al., 2013; Corrigan, Markowitz, Watson, Rowan, & Kubiak, 2003; Phelan et al., 2002). Moreover, causal attributions have important implications for perceptions of self-efficacy or confidence in one’s ability to improve their health status (e.g., Dar-Nimord, Cheung, Ruby, & Heine, 2014; Persky & Eccleston, 2011) and, in turn, health-related behavior (Knerr, Bowen, Beresford, & Wang, 2014). For example, biological attributions for one’s own obesity are negatively associated with self-blame, but also with self-efficacy and weight-malleability beliefs (Pearl & Lebowitz, 2014) which is associated with reduced fruit and vegetable intake and physical activity (Knerr et al., 2014).

The Obesity Stigma Asymmetry Model specifically lays out these multiple routes through which high versus low-control attributions can lead to higher versus lower levels of stigma (Hoyt, Burnette, Auster-Gussman, Blodorn, & Major, 2017). This model proposes both positive and negative outcomes of low-control attributions via different psychological processes. Here, low-control attributions (such as genetics) decrease anti-fat attitudes by reducing blame towards individuals for causing their condition, but also increase these attitudes by strengthening beliefs about obesity’s unchangeable and fixed nature. Conversely, high-control causal factors (such as lifestyle behavior) increase anti-fat attitudes by strengthening blame, but also decrease these attitudes by reducing beliefs that obesity is unchangeable. Although this model was developed in the context of obesity it is relevant for understanding processes related to diabetes.

In the context of diabetes, literature about the role of causal attributions on beliefs and judgments is sparse. Two previous papers have reported that behavioral causal attributions for both T1D and T2D were negatively correlated with ratings of favorability toward affected individuals (Rose et al., 2019; Anderson-Lister & Treharne, 2014). Schabert and colleagues also point to causal attributions as a driver of internalized stigma for people with T2D and note that this stigma is associated with self-blame and negative stereotyping of other affected individuals (Schabert, Browne, Mosely, & Speight, 2013). The present investigation builds on this work to systematically examine the associations of causal attributions with internalized stigma and self-blame for T1D and T2D. From here, we go on to investigate the pathways through which such stigma and blame could influence patient health and well-being among both disease types.

Blame, Stigma, and Downstream Consequences

There is a strong body of literature demonstrating that self-blame and stigma can have deleterious consequences for individuals who experience them. The negative consequences of having a stigmatized identity have been clearly laid out in the context of weight, with research indicating that weight stigma can impair health by increasing stress, negative emotions, physiological reactivity, and also lead to poorer self-regulation and lifestyle behavior (Major, Tomiyama & Hunger, 2018; Hunger, Major, Blodorn, & Miller, 2015). Similar negative relationships between stigma and health have also been noted in many health conditions such as HIV/AIDS, inflammatory bowel disease, and mental illness (Rueda et al., 2016; Sickel, Seacat, & Nabors, 2014; Taft, Keefer, Leonhard, & Nealon-Woods, 2009).

In the context of diabetes, a review by Schabert and colleagues (2013) points to several negative consequences of stigma, including psychological distress and sub-optimal self-management of the condition, which culminate in poorer clinical outcomes. For T1D, qualitative research has posited a pathway between stigmatizing aspects of T1D and deleterious coping strategies among affected individuals (Nishio & Chujo, 2017). For example, stigma experiences such as feeling socially excluded result in coping strategies like disease concealment and limiting of social opportunities, which can exacerbate negative health and quality of life outcomes. Other qualitative work in T1D cites a host of negative consequences of stigma, including increased emotional distress and poor disease management (Browne, Ventura, Mosely, & Speight, 2014). Research among younger affected individuals found that self-blame is related to poorer quality of life and is one of the most important predictors of psychological maladjustment (Kraaij & Garnefski, 2015).

In the context of T2D, qualitative research has linked stigma experiences with concealment of diabetes status and with psychological distress (Browne, Ventura, Mosely, & Speight, 2013). Such distress has further been linked to lower treatment adherence rates in this population (Gonzalez, Kane, Binko, Shapira, & Hoogendoorn, 2016). Research has further shown internalized stigma among individuals affected by T2D, which is associated with poorer attitudes towards self-care behavior, lower self-efficacy for illness management, and greater diabetes complications (Kato et al., 2017; 2016). Thus, the literature linking both T1D and T2D stigma and self-blame with poorer self-care and disease management as well as negative mental and physical health outcomes is small but consistent.

Current Study

Although the literature is suggestive of potential pathways linking diabetes attributions with health and well-being outcomes, there are gaps in understanding how the causal attributions individuals make about their diabetes associate with self-blame and internalized stigma. There are also gaps in understanding paths to downstream health and well-being consequences stemming from these negative self-judgments. The current study investigates these pathways for T1D and T2D individually; however, inclusion of both diabetes types allows for side-by-side comparison of whether and how different perceived etiologies might result in differential relationships among variables. For example, although affected individuals make genetic attributions for both T1D and T2D, these attributions may be more influential on stigma and self-blame among those affected by T2D in that they might buffer against negative effects of the strong behavioral causal beliefs for this diabetes type. Gaining a clearer understanding of these pathways initiating from causal attributions could be useful in creating interventions to improve negative cognitions associated with causal beliefs and, in turn, improve downstream outcomes such as self-care and overall life satisfaction. In addition, the ability to compare these processes among individuals with T1D and T2D may prove instructive for theory building.

Research questions for the current analysis were as follows:

1. Do genetic and behavioral causal attributions that affected individuals make for their T1D versus their T2D relate differently to self-blame and internalized stigma?

2. Is there support for a model wherein causal attributions for one’s diabetes initiate a pathway, operating through self-blame and internalized stigma, on outcomes including self-care, diabetes symptoms, and life satisfaction?

We specifically propose that causal attributions will influence self-blame and internalized stigma, which will influence participants’ levels of diabetes self-care and will go on to influence their symptom reporting, which will feed into life satisfaction. We chose to test this particular order because negative self-oriented beliefs have been linked with suboptimal self-care in past work (Schabert et al., 2013), and one’s disease management behaviors logically feed into the experience of symptoms. Finally, experience of disease symptoms is one element of many that can influence one’s larger life satisfaction (Al-Windi, 2005). Path models will be developed for T1D and T2D separately so that support for the proposed path can be individually evaluated for each disease.

Method

Participants

Data for this report comes from the Diabetes Identity, Attributions, and Health Study (DINAH) (Costabile, Boland, & Persky, 2019; Rose et al., 2019), which collected cross-sectional data on individuals diagnosed with either T1D or T2D. Participants in the diabetes-affected arms of the DINAH study included 192 survey respondents affected by T1D and 207 affected by T2D. Most participants were recruited through the ResearchMatch platform where they had previously indicated their diabetes status when volunteering to be contacted for research studies. Additional participants were also recruited through postings in diabetes-specific Facebook groups. Participants were compensated $10 by gift card for completing the survey. This study was determined to be IRB-exempt by the appropriate office at the National Institutes of Health.

Procedure

Following recruitment into the study, participants completed an online survey administered via SurveyMonkey. No personally identifying information was collected. Participants in the current analysis were asked to indicate their diabetes diagnosis as T1D, T2D, or unsure. Only those who indicated diagnosis with T1 or T2D were included in data collection. In order to be retained in the data set, participants were required to pass a knowledge check indicating that they understood the difference between T1 and T2D (see Rose et al., 2019), and were also asked again at the end of the survey about their diabetes status and whether there is any reason why their data should not be used (upon reassurance that they would receive payment regardless). The final data set consisted of 177 individuals affected by T1D and 186 individuals affected by T2D.

Measures

Causal Attributions.

All respondents received items assessing causal attributions for each type of diabetes based on items from several other causal attribution scales, see Rose et al. (2019). The current analysis included items assessing genetic, dietary, and physical activity-oriented attributions for participants’ own diabetes type. Participants with T1D therefore indicated “the extent to which [they] agree or disagree that the following factors cause or contribute to a person’s risk for getting Type 1 diabetes sometime in his/her lifetime”, where factors included “dietary factors, regardless of weight”, “physical activity”, and “genetics”. For participants with T2D, the same items were included. Responses were collected on a 1 “strongly disagree” to 7 “strongly agree” scale. Dietary and physical activity factors were combined to provide an assessment of behavior causal attributions for both groups (the two items were correlated r = .83, p < .01 for T1D and r = .72, p < .01 for T2D).

Internalized Stigma and Self-blame.

All respondents received items assessing stigmatization of their own type of diabetes using the Universal Measures of Bias, negative attitudes and distancing subscales (Latner, O’Brien, Durso, Brinkman, & MacDonald, 2008). For example, participants with T1D were given the following instructions: “indicate the extent to which you agree or disagree with the following statements about people affected by Type 1 diabetes. Please think about your view of people with Type 1 diabetes in general, rather than any specific person you may know who is affected by the disease.” The stigma scale consisted of two subscales, negative attitudes (Cronbach’s alpha for T1D=.93, T2D=.89) and distancing (Cronbach’s alpha for T1D=.68, T2D=.77). A sample negative attitude item is “People with Type 1 diabetes tend toward bad behavior,” whereas a sample distancing item is “I would not want to have a person with Type 1 diabetes as a roommate.” All responses were indicated on a 1–7 scale from “strongly disagree” to “strongly agree”. The current analysis includes responses relevant to participants’ own type of diabetes. As such, this measure represents the extent to which participants endorse and have internalized stereotypes and biases against their own stigmatized group. Self-blame was measured with a single item: “I feel I am to blame for my diabetes,” answered on a 1–4 scale from “not at all true” to “completely true.”

Self-care.

Self-care was assessed using the Diabetes Self-Management Questionnaire (DSMQ) (Schmitt et al., 2013). Participants were asked to consider their self-care over the last 8 weeks and indicate the extent to which several statements applied to them. The response scale went from 0 “does not apply to me” to 3 “applies very much to me.” A score for self-care was calculated in accordance with scale scoring instructions. A sample item is, “I do regular physical activity to achieve optimal blood sugar levels”.

Symptom report.

Symptom reporting was assessed using a shortened version of the Diabetes Symptom Checklist-Revised (Arbuckle et al., 2009). Participants were asked to indicate how often they experienced a number of hyper- and hypoglycemia symptoms within the past four weeks. The scale ranged from 0 “not at all” to 3 “daily.”

Life satisfaction.

Life satisfaction was assessed using the Satisfaction with Life Scale (Diener, Emmons, Larsen, & Griffin, 1985) which consisted of 5 items (Cronbach’s alpha=.92). Responses ranged from 1 “strongly disagree” to 7 “strongly agree.” A sample item is, “The conditions of my life are excellent”.

Data Analysis

Descriptive statistics were calculated for demographics and other study variables, comparing the T1D and T2D groups. Moderation analyses and path analysis included BMI, age, and education as covariates given significant differences between diabetes types. For moderation analyses, three regression analyses were run, one for each outcome variable: 1) internalized stigma: negative attitude, 2) internalized stigma: distance, and 3) self-blame. For each analysis, diabetes type, behavioral attributions, genetic attributions, and the interaction of diabetes type with each type of attribution were entered as simultaneous predictors. The three analyses without the interaction term were also conducted and are available in the supplemental table.

To examine the potential downstream consequences of these moderation analyses, we used SPSS AMOS to develop a path model, again including BMI, age, and education as covariates. Consistent with the moderation analysis findings, separate models were conducted for individuals affected by T1D versus T2D. As this path model is exploratory, we first constructed a fully-identified model and specified genetic attributions and behavioral attributions as exogenous variables, with internalized stigma: negative attitude, internalized stigma: distance, self-blame, self-care, symptoms, and life satisfaction all specified as endogenous variables. Next, we removed paths to examine proposed relations among variables, with remaining paths leading from attributions to negative self-judgments (internalized stigma: negative attitude, internalized stigma: distance, and self-blame). Based on previous literature, we proposed that these negative judgments would affect self-care behavior (Kato et al., 2016; Schabert et al., 2013), which would then be associated with symptoms (Weinger, Butler, Welch, & La Greca, 2005), which would then be associated with life satisfaction (Arditi, Zanchi, & Peytremann-Bridevaux, 2019; Imayama, Plotnikoff, Courneya, & Johnson, 2011). The model fit indices compare the proposed downstream consequences of the causal attribution model with a fully identified model. Hu & Bentler’s (1999) recommendations of interpreting fit were used, RMSEA < .06, CFI > .95, and a non-significant chi-square test, as indices as good model fit.

Results

Demographics

Demographic characteristics of participants are available in Table 1. Participants affected by type 1 diabetes (T1D) and type 2 diabetes (T2D) differed by age, BMI, education level, and race. The two groups did not differ by gender composition.

Table 1 –

Demographic characteristics of individuals affected by T1D and T2D

	T1D	T2D	Comparison
Age	40.39 (14.30)	54.01 (11.47)	F(1,370)=103.09, p<.0001
BMI	26.6 (5.79)	33.54 (9.12)	F(1,369)=74.54, p<.0001
College graduate	126 (71%)	102 (55%)	X2=9.47, p=.002
White race	160 (90%)	149 (80%)	X2=6.56, p=.010
Female	138 (78%)	135 (72%)	X2=1.29, p=.26

Between-group means and differences

Means for all variables are available in Table 2. Individuals affected by T1D and T2D did not significantly differ in their genetic causal attributions for their own condition. In contrast, individuals with T2D made far greater behavioral attributions for their condition than individuals with T1D. Concerning negative attitudes, individuals with T2D reported more negative attitudes about their condition than those with T1D. In contrast, there was no difference in distancing attitudes between individuals with T1D and T2D about their own condition.

Table 2 –

Variable means and standard deviations by individuals affected by type 1 diabetes (T1D) and type 2 diabetes (T2D)

	T1D	T2D	Comparison
Genetic attribution for own diabetes type	5.74 (1.54)	5.87 (1.36)	F(1,356)=3.41, p=.066
Behavior attribution for own diabetes type	1.86 (1.43)	5.73 (1.33)	F(1,356)=366.87, p<.0001
Internalized stigma: negative attitudes subscale	1.97 (1.12)	3.99 (.84)	F(1,356)=299.46, p<.0001
Internalized stigma: distancing subscale	1.78 (.92)	1.90 (1.06)	F(1,356)=1.54, p=.22
Self-blame for diabetes	1.21 (.54)	2.48 (1.03)	F(1, 363)=138.64, p<.0001
Self-care	2.27 (.48)	2.15 (.56)	F(1,364)=7.08, p=.008
Symptom report	2.36 (1.43)	2.39 (1.56)	F(1,364)=.94, p=.33
Life satisfaction	4.71 (1.50)	4.18 (1.52)	F(1,364)=1.23, p=.27

Additionally, individuals affected by T1D expressed less self-blame and higher levels of self-care than individuals affected by T2D. Individuals with T1D versus T2D did not differ in their levels of symptom report or their reported life satisfaction. See Table 2.

Moderation by diabetes type on stigma and self-blame outcomes

The regression model predicting internalized stigma: negative attitudes revealed a main effect of diabetes type (B=2.82, p<.0001), a main effect of behavioral causal attributions for one’s own diabetes type (B=.30, p=.012), and an interaction between diabetes type and behavioral attributions (B=−.15, p=.037) such that negative attitudes were higher among those with T1D when behavioral attributions were higher, while negative attitudes did not depend upon behavioral attributions among those with T2D. Genetic causal attributions were not a significant predictor nor was there a genetic attribution-by-diabetes type interaction.

Internalized stigma: distancing was predicted by a main effect of diabetes type (B=2.14, p<.0001), a main effect of genetic causal attributions for one’s own diabetes type (B=.36, p=.001), and a diabetes type by genetic attribution interaction (B=−.28, p<.0001) such that distancing was low among those with T1D regardless of genetic attributions, while distancing was higher when genetic attributions were lower among those with T2D. Behavioral causal beliefs were not a significant predictor nor was there a behavioral attribution-by-diabetes type interaction.

Finally, self-blame was predicted by a main effect of diabetes type (B=1.63, p<.0001), a main effect of genetic causal attributions for one’s own diabetes type (B=.22, p=.018), and a diabetes type by genetic attributions interaction (B=−.18, p=.003) such that self-blame was relatively low among those with T1D regardless of genetic attributions, while self-blame was lower when genetic attributions were high among those with T2D. Behavioral causal attributions were not a significant predictor nor was there a behavioral attribution-by-diabetes type interaction. See Figure 1 for visualization of the interactions. Regression models without interaction terms are available in the supplemental table.

Figure 1:

Visualizations of interactions predicting internalized stigma and self-blame outcomes

Type 1 diabetes path model

See Figure 2 for standardized beta coefficients from the fully identified path model. For simplicity, non-significant paths are not presented in the figure, although they were specified in the tested model. Full path coefficients are reported in the supplemental table. Results demonstrated that, for T1D, genetic causal attributions were positively related to internalized stigma: distancing and also positively associated with self-blame, though this latter relationship did not reach statistical significance. Behavioral causal attributions were positively related to internalized stigma: negative attitudes and self-blame. From there, self-blame was negatively associated with self-care, which was negatively associated with symptoms and positively related to life satisfaction. Symptoms were also directly negatively associated with life satisfaction. To probe the relationships more carefully, we then omitted the unpredicted paths from the model (e.g., from causal attributions directly to downstream consequences). After these paths were removed, model fit was good, Χ² (13) = 20.742, p =.078, CFI = .963, RMSEA = .057, indicating that these paths did not contribute meaningfully to the model and providing further evidence for a proposed downstream consequence of causal attributions model.

Figure 2:

Path model of relationships between causal attributions, internalized stigma types, self-blame, self-care, symptoms, and life satisfaction for Type 1 Diabetes. Standardized betas are presented from the fully identified model. Only statistically significant relationships are presented in the figure.

†=p<.10, *= p<.05, **=p<.01

Type 2 diabetes path model

See Figure 3 for path model. Full path coefficients are reported in supplemental table Results of the fully identified path model demonstrated that, for T2D, unlike in the previous T1D model, genetic causal attributions were negatively related to internalized stigma: distancing, and self-blame. Behavioral causal attributions were positively related to self-blame. From there, self-blame and distancing were both negatively associated with self-care. Self-care was negatively associated with symptoms, while genetic causal attributions, negative attitudes, distance, and self-blame were all positively related to symptoms. Symptoms were also negatively related to life satisfaction. To probe the relationships, we then omitted the unpredicted paths from the model (e.g., from causal attributions directly to downstream consequences). After these paths were removed, model fit was good, Χ² (13) = 19.850, p =.099, CFI = .956, RMSEA = .053, indicating that these paths did not contribute meaningfully to the model providing further evidence for a proposed downstream consequence of causal attributions model.

Figure 3:

Path model of relationships between causal attributions, internalized stigma: negative attitudes, self-blame, self-care, symptoms, and life satisfaction for Type 2 Diabetes. Standardized betas are presented from the fully identified model. Only statistically significant relationships are presented in the figure.

†=p<.10, *= p<.05, **=p<.01

Discussion

In exploring the influence of causal attributions for individuals with T1D and T2D side-by-side, a clear pattern emerges as to how these beliefs have different implications as a function of diabetes type. The relationships early on in our proposed path model, from causal attributions to internalized stigma and self-blame, go on to a cascade of behaviors and outcomes that have important implications for health and life satisfaction. Results are suggestive of the possibility that by attending more closely to how diabetes causal factors are presented and discussed (such as in public health and healthcare settings), we may be able to influence these basic beliefs and, in turn, improve health and adaptation downstream.

Consider first T2D. Here, the relationship between genetic attributions and stigma is in line with conceptions of genetic attributions as a route for bias reduction. Indeed, genetic attributions were associated with lower levels of self-blame and internalized stigma. Making behavioral attributions for one’s T2D, on the other hand, was linked with increased self-blame and negative attitudes toward one’s condition in the current study. These findings echo a long history of work in the domain of weight and obesity, wherein attributing one’s weight status to genetics serves to alleviate blame and stigma while behavioral attributions serve to reinforce it (Khan et al., 2017; Kvaale et al., 2017; Lebowitz & Appelbaum, 2017; Hilbert, 2016; Pearl & Lebowitz, 2014; Persky & Eccleston, 2011; Bennett et al., 2008; Corrigan et al., 2003; Phelan et al., 2002; Crandall, 1994; Weiner, 1993; Weiner et al., 1988). This is unsurprising given T2D’s direct link with obesity as a predisposing factor (Schnurr et al., 2020), and that both conditions have sizeable and salient behavioral risk factors (National Institute of Diabetes and Digestive Kidney Diseases, 2016). It is yet to be determined how these patterns of causal attributions in T2D relate to motivation for health-promoting behaviors. This will be particularly important for attributions made for disease progression rather than onset. It is interesting to note that the type of stigma that contributed to variance in the path model was distancing from affected individuals, while negative attitudes were unrelated to attributions. The reasons for this are unclear, but it may reflect a desire to preserve a positive self-concept by distancing from the stigmatized ingroup of individuals with T2D (Ellemers, van Knippenberg, & Wilke, 1990).

In contrast, for T1D, the association between genetic attributions and self-blame falls in the opposite direction. Here, those with higher genetic causal attributions reported more self-blame and more negative attitudes toward affected individuals. This is likely because T1D is less frequently considered to have personal behavior causes. Therefore, genetic causal beliefs cannot counteract or buffer against perceptions associated with behavioral causes that might otherwise be a source of blame and stigma. Relieved of its role in reducing conceptions of personal fault, genetic causal attributions may be considered through a more essentialist lens. Here, genetic causes can indicate that the disease is deep-seated and inseparable from the individual, as their DNA is flawed or problematic (Dar-Nimrod & Heine, 2011). This pattern of linkage between genetic causal attributions and increased stigmatization is also seen among other health conditions that are rarely considered to have a behavioral cause (e.g., mental illness; Lebowitz & Appelbaum, 2017; Kvaale et al., 2013; Bennett et al., 2008; Phelan et al., 2002). As such, the very meaning of a genetic attribution itself may change in tune with disease characteristics.

Overall, this study demonstrates that these variations on perceived genetic causal attributions and stigma can have important downstream correlates. Here, the stigma and self-blame that follow from causal attributions are linked with important outcomes related to patients’ self-care, symptom reporting, and ultimately life satisfaction. This is consistent with the notion that causal attributions can shape the mindsets with which one approaches their own illness, and these mindsets are exceedingly powerful predictors of psychological and health outcomes (Burnette, Hoyt, & Orvidas, 2017). It may also be that individuals with high levels of self-blame or stigma have a poorer self-image or self-concept and as such are less motivated to care for the self (Kato et al., 2016; Fung, Tsang, & Corrigan, 2008). As the current study was cross-sectional, there may also be other routes or patterns explaining the relationships among variables. For example, if patients believe that their poor self-care behaviors in the past might have contributed to their diagnosis in addition to their current disease state, they may exhibit more self-blame for their disease onset.

Of note is that for patients affected by T2D, the model identified several direct effects to reported symptoms that did not flow through self-care. These include genetic causal attributions, self-blame, and internalized stigma: distancing. For T1D, there was a similar, direct influence on symptom reports from genetic causal attributions and internalized stigma: distancing. Although these patterns can only be postulated and not be definitively delineated, this is consistent with findings in other diseases wherein experienced stigma and self-blame have direct, negative effects on health (Link & Phelan, 2006).

There are limitations of the current analysis. Most importantly, as noted above, this study was cross-sectional and, as such, causal relationships cannot be identified. Path analysis can help infer potential relationships, but these will need to be tested more definitively using other study designs. In addition, the sample in this study was a convenience sample, wherein participants were more likely to be interested in research participation and/or relating to other affected individuals through social and advocacy groups. The sample size was also relatively small when stratified by diabetes type and future work should examine these questions in larger samples. Finally, the measure of internalized stigma was general and broad. We used a universal measure of bias to encompass the many differences between T1D and T2D. Although distancing measures have been used as a proxy for self-stigma in the mental illness literature (e.g., Uchino, Maeda, & Uchimura, 2012), it has not previously been used in the diabetes literature to our knowledge. Participants reported relatively low levels of stigmatizing attitudes. More tailored stigma scales might reveal nuances in the particular stigmatizing attitudes at play in these processes.

In all, this study examined two diseases that share both a name, elements of their disease course, and genetic etiology. Yet, findings indicated that affected individuals attach very different meaning to those etiological attributions. These patterns suggest possible points for intervention even at this early stage. The negative tenor of genetic causal beliefs among individuals with T1D suggests a possible opportunity to communicate to affected individuals that attributing their disease to genetic factors does not indicate a weakness or a failing. As attempts to better elucidate the origins and causes of T1D continue, it will be important to be aware of how causal routes are presented. Behavioral causal attributions were also associated with stigma and self-blame, so it is beneficial for future work to elucidate the specific behaviors that make up these attributions in T1D as this may provide an opportunity to correct misconceptions. Additionally, among those with T2D, there may be an opportunity to reduce self-blame and stigma by communicating the genetic causal factors that exist alongside and in interaction with environmental and behavioral factors. These efforts seem to have the potential to improve patient health and life satisfaction through the routes described here.