Depression and the medicalization of sadness: conceptualization and recommended help-seeking.

BACKGROUND: Critiques of the validity of the DSM diagnostic criteria for depressive disorder argue that it fails to differentiate between abnormal sadness due to internal dysfunction or depression (sadness without an identifiable cause), and normal sadness (sadness with a clear cause). AIMS: and
METHODS: A population survey was undertaken in adult education centres in Spain aiming to explore beliefs about depression and normal sadness. Two hypothetical case vignettes portrayed individuals experiencing deep sadness, both fulfilling criteria for major depressive disorder (DSM-IV), one with a clear cause, the other without an identifiable cause. Three hundred and forty-four (344) questionnaires were obtained (95% response rate).
RESULTS: Participants statistically significantly differentiated between the sadness-with-cause vignette, seen more frequently as a normal response, while the one without a cause was seen as pathological. Help-seeking behaviour recommendations followed this distinction: a medical option was statistically significantly more common when there was no cause for sadness. Socio-cultural variation in how people understand and deal with sadness was also found.
CONCLUSIONS: This study emphasizes the importance of taking into account the context in which depressive symptoms occur as it seems that the absence of an appropriate context is what makes people conceptualize them as abnormal. It also raises questions about the lack of face validity of the current diagnostic classification for depressive disorder that exclusively uses descriptive criteria.

In memory of Professor Peter Goldie, philosopher (1946–2011)

Introduction

Medicalization of normal sadness as depressive disorder

According to the World Health Organization (WHO), depression will be the second biggest disease burden by 2020 (Murray & Lopez, 1996). However, some of this increase may be due to misdiagnosis. Thus, a decline in the assessment of patients’ personal experiences and their cultural and social contexts (e.g. Andreasen, 2007; Dalal & Sivakumar, 2009; Jadhav & Littlewood, 1994) is likely to produce diagnoses overly based on symptoms. It is being argued that the current diagnostic criteria for diagnosing depressive disorder (since the DSM-III, in 1980) do not differentiate between abnormal sadness due to internal dysfunction or depression (sadness without an identifiable cause) and normal sadness (sadness with a clear cause) (Horwitz & Wakefield, 2007). While some state that the current DSM definition of depression wrongly encompasses a natural reaction to life events rather than a mental disorder (Horwitz & Wakefield, 2007; Parker, 2007), others defend the diagnostic criteria, raising concerns about people still being undiagnosed who are missing out on treatment (Hickie, 2007; Pies, 2009; Royal College of Psychiatrists, 1992).

There is an important exclusion clause in the diagnostic criteria for depressive disorder: people presenting with depressive symptoms are exempt from being diagnosed if they are suffering due to the death of a loved one. In this instance, attention is paid to the context of the symptoms being seen as a normal human response. Horwitz and Wakefield (2007) propose that besides bereavement, there are many other types of loss and adverse events that can trigger deep but normal sadness that also need to be taken into account. Moreover, the DSM also recognizes the possibility of grief becoming abnormal – complicated grief – when the symptoms become too severe and persistent; it is precisely the evaluation of the severity of the symptoms that decides the normality or abnormality of the response and that indicates that a diagnosis may be warranted. The same process could be used when assessing deep sadness due to other causes and not just when it is caused by bereavement.

Culture, religion and depression

Generally speaking, individuals seek to make sense of events and experiences, particularly when the events are unusual or have significance in terms of their effect on self or others. The significance or meaning that individuals give to this or that event is often contextualized by its occurrence in the midst of antecedent objective events. Thus, it is generally considered acceptable that people who have experienced a defined loss (e.g. bereavement, relationship breakdown) may become withdrawn, silent and tearful. Conversely, we might consider it very odd if people in similar circumstances were to act in a cheerful, upbeat manner. Experience and cultural factors provide the benchmark for what is understandable or acceptable behaviour and influence attitudes and beliefs about illness, determining help-seeking behaviours (Kleinman, 1981). It is worth noting that phenomena viewed as depression in western countries are rarely understood as a mental illness in many developing countries (Kleinman & Good, 1985). When looking across cultures, an experience of everyday distress which involves the loss of something essential of oneself emerges (e.g. ‘soul loss’) (Littlewood, 2002).

There is evidence that some aspects of religion are associated with positive mental health (Dein, 2006; Kang & Romo, 2010; Koenig, McCullogh, & Larson, 2001; Levin, Chatter, Ellison, & Taylor, 1996). Various literature reviews of religion and depression suggest that religion may be a protective factor for depression. Several studies undertaken in the USA and Canada show an association between attending church and a lower likelihood of suffering from depression. There is also evidence that being religious may have a positive effect on the outcome of depression, increasing the speed of recovery (Koenig, 1997; Koenig et al., 2001). Religious communities and faith-based organizations provide social support and religious belief systems offer meaning and hope about the future as well as providing the individual with a means of coping in the midst of suffering (Grosse-Holforth et al., 1996; Koenig et al., 2001; Krause, 1995; Tix & Frazier, 1998). Koenig and colleagues (1992) showed, in a study examining the relationship between depression and the use of religion as a coping behaviour in hospitalized medically ill men, that the only characteristic that predicted lower rates of depression six months later was the extent to which patients relied on their religious faith to cope. Another large study by the same author found that people who attended church frequently had lower rates of depression (Koenig, George, Meador, Blazer, & Dyck, 1994).

Those who are religiously involved have been found to have more positive attitudes towards life and to experience greater life satisfaction, which in turn may contribute to psychological well-being (Koenig et al., 1994). An American study showed that relying on religion as a coping strategy was associated with a high level of self-esteem (Krause, 1995).

Although much of the literature suggests an overall positive effect of religion on mental health, many authors have warned against the putative harmful effects of religion such as sectarianism, guilt, obsessionality and mental inflexibility (Koenig, 1997). The measurement of religion used in some studies may also be problematic as, for example, church attendance does not necessarily reflect genuine religious faith. Another area of concern is the possibility of a delay in seeking help for mental health problems, leading to a worsening of their prognosis, due to an excessive reliance on religious rituals and prayer (Dein, 2006). Religious people suffering from mental health problems may prefer to seek spiritual care and the advice of the clergy rather than advice from secular medical professionals (Hatfield, Mohamad, Rahim, & Tanweer, 1996; McCabe & Priebe, 2004; Mitchell & Baker, 2000; Wang, Berglund, & Kessler, 2003; Weaver, Flannelly, Flannelly, & Oppenheimer, 2003).

Aims

We aimed to explore beliefs about depression and normal sadness among a sample of people in Spain. We were interested in: (1) how lay people conceptualized sadness with a clear cause and sadness without any obvious cause; (2) whether the presence or absence of obvious causation influenced help-seeking; and (3) to what extent their conceptualizations were associated with other socio-cultural factors. We also wanted to investigate the role religion (measured as church engagement) played in how people understood the characters’ experiences of sadness and how they sought to resolve or alleviate it, being particularly interested in finding out whether the clergy was recommended as a source of help in times of sadness.

Three hypotheses were formulated. The first two were based on the current critique of depression presented above and the third on the studies looking at help-seeking behaviour of religious people suffering from mental health problems:

Sadness without a perceived cause will be more frequently conceptualized as an illness, whereas sadness with a cause will be understood more in terms of a misfortune, as a response to the vicissitudes of life, but not as pathological.

Help-seeking behaviour to deal with sadness will be influenced by its conceptualization: those who understood sadness as an illness will be more likely to recommend seeking some form of medical attention.

Those participants with more frequent church attendance will more often propose seeking the advice of a priest to deal with sadness than those with more infrequent practice.

Method

Procedures and sampling

We approached four adult educational centres, associated or managed by Catholic organizations in the region of Valencia[1] (Spain). All four centres approached agreed to participate. Three centres offered courses to immigrants (mostly Central and South Americans) in and around the city of Valencia and the fourth was a theological college with 10 satellite centres in villages. Teachers and administrative personnel distributed self-completed questionnaires among the students.

The inclusion criteria included having Spanish as a first language and at least intermediate reading level. Ethical approval to undertake the study was granted by the Imperial College London Research Ethics Committee.

Measures

Our self-designed questionnaire included socio-demographic information and questions about religious affiliation and frequency of attendance at a place of worship. Two hypothetical case vignettes followed, both portraying a deeply sad and distressed person. One vignette depicted a woman experiencing sadness with an apparent cause for this suffering (her daughter’s diagnosis of a life-threatening illness); the other described a man undergoing sadness without an apparent cause. In both cases, the character portrayed fulfilled the criteria for major depressive disorder (DSM-IV) (American Psychiatric Association, 2005) and for severe depressive episode (ICD-10) (World Health Organization, 1992). After each vignette, participants were asked to indicate on a multiple-choice questionnaire what might be happening to the character and what he/she ought to do to get better. The options were developed following a pilot questionnaire that was tested with 30 volunteers including both Spanish nationals and Central and South Americans (these responses were not included in the results). The first question asked participants what they thought was happening to the character and offered the following three options: misfortune, reaction to stress or illness. In addition, they could specify the type of illness: mental, physical or both. A space was provided to allow participants a written answer describing what they thought the character was experiencing. A second question to elicit what the character ought to do to get better followed, asking participants to choose from the following help-seeking options: (1) medical attention (in general); (2) a general practitioner; (3) a psychiatrist or psychologist; (4) relatives/friends; (5) a priest; (6) a folk healer or a ‘wise person’; and (7) self-help (‘getting better by yourself’).

Analysis

We undertook simple descriptive statistics. We then used Pearson χ2 tests with Yates’ correction in order to investigate possible associations between cultural and socio-demographic variables, and the conceptualization of sadness and recommended help-seeking. The following socio-demographic variables were studied: age, gender, education level, employment, ethnicity, level of religious practice, living arrangements, civil status, having children and legal status. These variables were divided into groups for further analysis, for example education level: school (primary/secondary), further education, and university. Fisher’s exact tests were used where appropriate. Finally, binary logistic regression was carried out to investigate which variables might predict the conceptualization of sadness and help-seeking. This was carried out as stepwise forward regression, by adding single predictors. They were identified on the basis of the strength of the association with the specified outcome variable (Field, 2009). All analyses were conducted using the Statistical Package for Social Sciences (SPSS, 2003) for Windows and a 95% (p < .05) statistical significance level was applied.

Results

In this section, the sample characteristics and the main findings of the study are presented. Due to space constraints, not all the results have been displayed in tables, therefore statistical details are given in full only when the findings are not included in the tables.

Sample characteristics

Three hundred and forty-four (344) out of 362 people participated in the study, providing a 95% response rate. The majority of participants were Spanish (80.5%) and the rest were Central and South American (19.5%). The age range was 18–83 years (M = 48.7, SD = 16.1). Most were women (77.2%), married or cohabiting (60.7%) and had children (54.7%).

The participants were predominantly Catholic (95.0%); 3% described themselves as Christian without specifying denomination and a very small minority did not have a religious affiliation (2.0%). Most attended church frequently (80% attending daily or weekly); infrequent religious attendance (monthly and less than monthly) was reported in 11.5%; and 8.6% never attended a place of worship or said that they did not follow a religion.

While 35.0% held a university degree, 37.3% attended only primary or secondary school; the rest completed further education after leaving school (27.7%). Over half of the sample were not in paid employment (pensioners, students or housewives) (54.3%). The majority owned their homes (77.2%); 15.2% were in rented accommodation, with the remaining 7.6% living in local authority housing or homeless (the latter were in fact staying with family or friends). In terms of their legal status, the majority were Spanish nationals or had ‘leave to remain’ (89.2%); the rest were undocumented migrants (9.6%).

Conceptualization of sadness

Almost 70% of participants conceptualized sadness without a perceived cause as an illness, mostly of a mental type (depression in particular). Conversely, a similar proportion of people understood sadness associated with a cause as a misfortune, confirming hypothesis 1 (Table 1).

Table 1.

Responses to the vignettes regarding conceptualization and help-seeking of sadness with and without cause (frequencies, percentages).

Responses (N = 344)	Sadness without cause		Sadness with cause
	n	%	n	%
Conceptualization as
Misfortune	34	9.9	232	67.4
Reaction to stress	149	43.3	80	23.3
Illness [a]	233	67.7	131	38.1
Mental	127/233	54.5	54/131	41.2
Physical	17/233	7.3	24/131	18.3
Both	89/233	38.2	53/131	40.5
Names of illness given[b]
Depression	126/172	73.3	39/89	43.8
Stress	9/172	5.2	4/89	4.5
Sadness[c]	7/172	4.1	18/89	20.2
Help-seeking
Medical attention	162	47.1	96	27.9
Relatives/friends	235	68.3	231	67.2
Get better by yourself	132	38.4	92	26.7
Folk healer/wise person	55	16.0	59	17.2
Priest	167	48.5	23	68.6
General practitioner	85	24.7	77	22.4
Psychologist/psychiatrist	187	54.4	129	37.5

For the type of illness, percentages were calculated using as a total the number of participants who chose illness as an answer: without cause (n = 233) and with cause (n = 131).

For the names given to the illness, percentages were calculated using as a total the number of participants who wrote down an answer: without cause (n = 172) and with cause (n = 89).

Other words used besides ‘sadness’ were ‘sorrow’ and ‘grief’ (in Spanish: tristeza, pesar, pena).

Table 2 shows statistically significant associations between some socio-demographic variables and conceptualizing sadness as illness in both vignettes: people older than 65 years more frequently defined the problem as an illness than their younger counterparts; people with a lower level of education (school education only) were more likely to consider it an illness; and those unemployed more often understood it as pathological than those employed (Table 2). Furthermore, using logistic regression, these last two variables, employment and education, were statistically significant predictors for understanding sadness with cause as an illness (Table 3).

Table 2.

Variables and help-seeking associated with the conceptualization of sadness as an illness.

Conceptualization of sadness as an illness
Variables and help-seeking	Sadness without cause		Sadness with cause
	n	%	n	%
Age (years)	χ 2 = 11.06, 3 df*		χ 2 = 19.83, 3 df***
18–30 (n = 64)	42	65.6	20	31.3
31–45 (n = 78)	50	64.1	26	33.3
46–65 (n = 140)	88	62.9	46	32.9
> 65 (n = 62)	53	85.5	39	62.9
Education	χ 2 = 7.43, 2 df*		χ 2 = 20.65, 2 df***
School (primary/secondary) (n = 128)	98	76.6	68	53.1
Further education (n = 95)	59	62.1	30	31.6
University (n = 120)	75	62.5	32	26.7
Employment	χ 2 = 10.85, 1 df**		χ 2 = 20.99, 1 df***
Employed[a] (n = 156)	92	59.0	39	25.0
Unemployed[b] (n = 185)	140	75.7	91	49.2
Medical attention	χ 2 = 90.94, 1 df***		χ 2 = 90.55, 1 df***
Sadness conceptualized as illness	151/233	64.8	75/131	57.3
Sadness not conceptualized as illness	11/111	9.9	21/213	9.9
General practitioner	χ 2 = 17.02, 1 df***		χ 2 = 11.40, 1 df**
Sadness conceptualized as illness	73/233	31.3	42/131	32.1
Sadness not conceptualized as illness	12/111	10.8	35/213	16.4
Psychiatrist/psychologist	χ 2 = 31.76, 1 df***		χ 2 = 3.26, 1 df
Sadness conceptualized as illness	151/233	64.8	57/131	43.5
Sadness not conceptualized as illness	36/111	32.4	72/213	33.8
Relatives/friends	χ 2 = 7.20, 1 df**		χ 2 = 0.49, 1 df
Sadness conceptualized as illness	170/233	73.0	85/131	64.9
Sadness not conceptualized as illness	65/111	58.6	146/213	68.5

Employed included full-time, part-time and self-employed.

Unemployed included retired, students and housewives.

p < .05, ** p < .01, *** p < .001

No statistically significant differences were found among the following variables: gender, ethnicity, level of religious practice, living arrangements, civil status, having children and legal status; nor for the following help-seeking behaviours: priest, folk healer/wise person and getting better by yourself.

Table 3.

Binary logistic regression analysis for variables predicting conceptualization of sadness as illness and medical attention.

	Conceptualization of sadness as illness[a]			Medical attention
		Sadness without cause	Sadness with cause		Sadness without cause		Sadness with cause
		OR, 95% CI	OR, 95% CI		OR, 95% CI		OR, 95% CI
Step 1	Employment	2.16, 1.36–3.44**	2.90, 1.83–4.61***	Age	1.60, 1.27–2.01***	Illness	12.24, 6.94–21.61***
Step 2	Employment	2.05, 1.28–3.29**	2.63, 1.64–4.23***	Age	1.59, 1.23–2.06***	Illness	11.63, 6.56–20.60***
	Age	1.94, 0.89–1.45	1.27, 1.00–1.60*	Illness	16.67, 8.39–33.12***	Age	1.22, 0.92–1.61
Step 3	Employment	1.87, 1.93–3.09*	2.15, 1.30–3.54**	Age	1.67, 1.28–2.18***	Illness	11.69, 6.50–21.03***
	Age	1.10, 0.86–1.41	1.18, 0.93–1.50	Illness	14.54, 7.24–29.18***	Age	1.23, 0.93–1.64
	Education	0.85, 0.63–1.15	0.69, 0.51–0.93*	Stress	0.56, 0.33–0.96*	Employment	1.02, 0.56–1.84

The same variables were added in steps 1, 2 and 3 for both vignettes (with and without cause) as they were found to be statistically significantly associated in both scenarios.

OR = odds ratio, 95% CI = 95% confidence interval for the odds ratio, lower – upper

p < .05, ** p < 0.01, *** p < .001

Illness (sadness without cause): R 2 = .03 (Cox & Snell), R 2 = .04 (Nagelkerke), Model χ 2 = 10.86, 1 df, p = .001 (step 1); R 2 = .03 (Cox & Snell), 0.05 (Nagelkerke), Model χ 2 = 11.93, 2 df, p = .003 (step 2); R 2 = .04 (Cox & Snell), R 2 = .05 (Nagelkerke), Model χ 2 = 12.69, 3 df, p = .005 (step 3).

Illness (sadness with cause): R 2 = .06 (Cox & Snell), R 2 = .08 (Nagelkerke), Model χ 2 = 21.44, 1 df, p = .000 (step 1); R 2 = .07 (Cox & Snell), R 2 = .09 (Nagelkerke), Model χ 2 = 25.44, 2 df, p = .000 (step 2); R 2 = .09 (Cox & Snell), R 2 = .12 (Nagelkerke), Model χ 2 = 30.72, 3 df, p = .000 (step 3).

Medical attention (sadness without cause): R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 = 17.43, 1 df, p = .000 (step 1); R 2 = .28 (Cox & Snell), R 2 = .38 (Nagelkerke), Model χ 2 =114.89, 2 df, p = .000 (step 2); R 2 = .29 (Cox & Snell), R 2 = .39 (Nagelkerke), Model χ 2 = 119.37, 3 df, p = .000 (step 3).

When in further steps, employment and level of religious practice were loaded, they did not reach a statistically significant value but age, conceptualization as illness and as stress remained statistically significant.

Medical attention (sadness with cause): R 2 = .23 (Cox & Snell), R 2 = .34 (Nagelkerke), Model χ 2 = 91.34, 1 df, p = .000 (step 1); R 2 = .24 (Cox & Snell), R 2 = .34 (Nagelkerke), Model χ 2 =93.32, 2 df, p = .000 (step 2); R 2 = .24 (Cox & Snell), R 2 = .35 (Nagelkerke), Model χ 2 = 93.70, 3 df, p = .000 (step 3).

When in step 4, education was loaded, it did not reach statistically significant value but conceptualization as illness remained statistically significant.

Married or cohabitating people were statistically significantly more likely to conceptualize sadness with cause as a misfortune than those single or widowed (72.0% vs 60.4% (χ 2 = 4.93, 1 df, p < .05)). Moreover, marital status was also a statistically significant predictor for conceptualizing sadness with cause as a misfortune (step 1, OR = 0.59, 95% CI = 0.37–0.94, p < .05).

Help-seeking for sadness

Medical help-seeking

Our second hypothesis was confirmed: the need for medical attention was higher for the scenario depicting sadness without cause than with cause (47.1% vs 27.9%). Similarly, seeking help from a psychiatrist or psychologist was more frequently selected for sadness without cause than with cause (54.4% vs 37.55%) (Table 1). Besides, those who understood sadness as an illness were statistically significantly more likely to recommend some form of medical attention than those who did not (without cause: 64.8% vs 9.9%; with cause: 57.3% vs 9.9%) (Table 2). Moreover, using logistic regression, conceptualizing sadness as an illness statistically significantly predicted recommending medical attention in both cases (Table 3).

Table 4 shows that some of the socio-demographic variables associated with the conceptualization of sadness as an illness were also found to be associated with seeking medical attention. Thus, those in the oldest age group (above 65 years of age) and those with no employment reported it the most in both scenarios; those with only school education also chose the option of seeking medical attention more often in the sadness-with-cause scenario.

Table 4.

Variables associated with medical help-seeking of sadness with and without cause.

Variables	Help-seeking
	Medical attention in general				General practitioner
	Sadness without cause		Sadness with cause		Sadness without cause		Sadness with cause
	n	%	n	%	n	%	n	%
Age (years)	χ 2 = 24.03, 3 df***		χ 2 = 13.62, 3 df**		χ 2 = 4.31, 3 df		χ 2 = 14.39, 3 df**
18–30 (n = 64)	22	34.4	14	21.9	12	18.8	10	15.6
31–45 (n = 78)	32	41.0	18	23.1	20	25.6	15	19.2
46–65 (n = 140)	62	44.3	35	25.0	32	22.9	27	19.3
> 65 (n = 62)	46	74.2	29	46.8	21	33.9	25	40.3
Religious practice	χ 2 = 9.16, 2 df*		Fisher’s exact = 1.80		χ 2 = 0.23, 2 df		Fisher’s exact = 3.86
Never/no religion (n = 29)	6	20.7	5	17.2	7	24.1	4	13.8
Infrequent[a] (n = 39)	20	51.3	10	25.6	11	28.2	5	12.8
Frequent[b] (n = 271)	135	49.8	79	29.2	67	24.7	67	24.7
Education	χ 2 = 1.563, 2 df		χ 2 = 8.00, 2 df*		χ 2 = 10.56, 2 df**		χ 2 = 3.52, 2 df
School (primary/secondary) (n = 128)	66	51.6	45	35.2	39	30.5	35	27.3
Further education (n = 95)	43	45.3	28	29.5	12	12.6	16	16.8
University (n = 120)	53	44.2	23	19.2	34	28.3	26	21.7
Employment	χ 2 = 10.18, 1 df**		χ 2 = 6.96, 1 df*		χ 2 = 1.25, 1 df		χ 2 = 4.57, 1 df*
Employed[c] (n = 156)	59	37.8	33	21.2	34	21.8	27	17.3
Unemployed[d] (n = 185)	102	55.1	63	34.1	50	27.0	50	27.0

Infrequent religious practice: monthly and less than monthly attendance to a place of worship.

Frequent religious practice: daily and weekly attendance to a place of worship.

Employed included full-time, part-time and self-employed.

Unemployed included retired, students and housewives.

No statistically significant differences were found among gender, ethnicity, living arrangements, civil status, having children and legal status.

No statistically significant differences were found for seeking the help of a psychologist/psychiatrist *P < .05; **P < .01; ***P < .001.

Non-medical help-seeking

Approximately 70% of participants recommended seeking help from relatives and friends for both scenarios. Consulting a priest was also commonly chosen: it was recorded as frequently as seeking the support of family and friends – almost 70% – for the sadness-with-cause vignette and in almost half of the participants for the sadness-without-cause vignette (Table 1).

Understandably, those who understood sadness with cause in terms of a misfortune were more likely to suggest that the character needed ‘to get better by themselves’ than those who did not (31.0% vs 17.9% (χ 2 = 6.69, 1 df, p < .05)). Furthermore, regression analysis showed that understanding sadness as a misfortune was a statistically significant predictor for suggesting that ‘getting better by oneself’ was what was needed to overcome sadness with cause (step 1, OR = 2.07, 95% CI = 1.18–3.62, p < .05).

Regarding ethnic variation, it was found that the Spanish group selected seeking help from friends and relatives more frequently than Central/South Americans in the sadness-without-cause scenario (41.2% vs 26.9% (χ 2 = 4.66, 1 df, p < .05)). Similarly, differences in legal status were found, with Spanish nationals and those with residence advising their help more than participants without residence (without cause: 40.7% vs 18.2% (χ 2 = 6.39, 1 df, p = .01); with cause: 69.4% vs 48.5% (χ 2 = 5.92, 1 df, p < .05)). Moreover, using logistic regression, legal status was found to be a statistically significant predictor for indicating this type of help in the sadness-with-cause vignette. Conceptualizing sadness as a misfortune and education beyond school level were also statistically significant predictors of recommending the help of family and friends in this vignette (Table 5).

Table 5.

Binary logistic regression analysis for predictors of seeking the help of relatives and friends, and of a priest.

	Relatives and friends’ help				Priest’s help
		Sadness without cause		Sadness with cause		Sadness without cause		Sadness with cause
		OR, 95% CI		OR, 95% CI		OR, 95% CI		OR, 95% CI
Step 1	Illness	1.91, 1.19–3.07**	Misfortune	2.73, 1.69–4.39***	Ethnicity	0.29, 0.16–0.52***	Ethnicity	0.35, 0.20–0.61***
Step 2	Illness	2.04, 1.26–3.32**	Misfortune	2.60, 1.61–4.21***	Ethnicity	0.28, 0.14–0.58**	Ethnicity	0.49, 0.25–0.95*
	Ethnicity	0.45, 0.26–0.79**	Education	1.52, 1.15–2.01**	Religious practice	0.95, 0.61–1.47	Religious practice	1.41, 0.92–2.16
Step 3	Illness	2.07, 1.27–3.39**	Misfortune	2.67, 1.63–4.35***	Ethnicity	0.36, 0.15–0.87*	Ethnicity	0.51, 0.26–1.01
	Ethnicity	0.53, 0.27–1.05	Education	1.49, 1.12–1.98**	Religious practice	1.01, 0.64–1.58	Religious practice	1.43, 0.93–2.21
	Religious practice	1.29, 0.84–1.99	Legal status	0.43, 0.20–0.90*	Legal status	0.69, 0.23–2.16	Misfortune	2.10, 1.28–3.44**

OR = odds ratio, 95% CI = 95% confidence interval for the odds ratio, lower – upper

p < .05, ** p < 0.01, *** p < .001

Relatives/friends (sadness without cause): R 2 = .02 (Cox & Snell), R 2 = .03 (Nagelkerke), Model χ 2 = 7.05, 1 df, p = .008 (step 1); R 2 = .04 (Cox & Snell), R 2 = .06 (Nagelkerke), Model χ 2 =14.63, 2 df, p = .001 (step 2); R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 = 17.61, 3 df, p = .001 (step 3).

Relatives/friends (sadness with cause): R 2 = .02 (Cox & Snell), R 2 = .03 (Nagelkerke), Model χ 2 = 7.05, 1 df, p = .008 (step 1); R 2 = .04 (Cox & Snell), R 2 = .06 (Nagelkerke), Model χ 2 =14.63, 2 df, p = .001 (step 2); R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 = 17.61, 3 df, p = .001 (step 3).

Priest (sadness without cause): R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 = 18.61, 1 df, p = .000 (step 1); R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 =17.42, 2 df, p = .000 (step 2); R 2 = .05 (Cox & Snell), R 2 = .07 (Nagelkerke), Model χ 2 = 17.35, 3 df, p = .001 (step 3).

Priest (sadness with cause): R 2 = .04 (Cox & Snell), R 2 = .05 (Nagelkerke), Model χ 2 = 13.68, 1 df, p = .000 (step 1); R 2 = .04 (Cox & Snell), R 2 = .06 (Nagelkerke), Model χ 2 =14.73, 2 df, p = .001 (step 2); R 2 = .07 (Cox & Snell), R 2 = .09 (Nagelkerke), Model χ 2 = 23.43, 3 df, p = .000 (step 3).

When in further steps, civil status, legal status and age were loaded, they did not reach statistically significant value but misfortune remained statistically significant.

Church engagement and help-seeking

When the whole (predominantly very religious) sample was confronted by the character undergoing sadness in the absence of a cause, recommending the help of a priest and medical attention reached very similar numbers (almost 50% for both sources of help) (Table 1). Moreover, those who never attended religious services or had no religious affiliation had approximately half the rate of recommending medical care than those who attended church to some degree (infrequent and frequent use) for this vignette (20.7% vs approximately 50%) (Table 4). Similarly, those without religious affiliation or who never attended church suggested less often resorting to family and friends than those who practised their religion (44.8% vs approximately 70% (χ 2 = 8.21, 2 df, p < .05)).

Help from the clergy

The overall percentages for selecting this answer were very high: almost half of the participants suggested it for the sadness-without-cause scenario and the majority (almost 70%) for the sadness-with-cause vignette (Table 1). Not surprisingly our third hypothesis was confirmed: those with frequent religious practice were statistically significantly more likely to choose the help of a priest when compared with participants with infrequent practice in both scenarios (without cause: 53.1% vs 20.5% (χ 2 = 13.24, 1 df, p < .001); with cause: 73.8% vs 51.3% (χ 2 = 7.33, 1 df, p < .01)). Those above 65 years old suggested seeking the help of a priest more often than younger participants when facing adversity (62.9% vs approximately 30% for the remaining age ranges which had similar percentages) (χ 2 = 11.63, 3 df, p < .001). This is understandable as there were statistically significant differences among age ranges for attending religious services: those above 65 years old had the highest level of church engagement (approximately 97% attended at least once a week vs approximately 44%–90% for the remaining age ranges) (χ 2 = 81.36, 6 df, p < .001). Nevertheless, no differences between genders were found in our study in the rates of advocating his help (nor in their frequency of church attendance).

Other statistically significant differences were found in both vignettes for seeking the help of a priest as follows: (1) those in local authority housing or who were homeless selected this help-seeking more often (without cause: local authority (57.7%), rented (26.9%), owned (52.3%) (χ 2 = 12.06, 2 df, p < .01); with cause: local authority (73.1%), rented (53.8%), owned (70.8%) (χ 2 = 6.084, 2 df, p < .05)); (2) Spanish nationals and those with residence advised it more than participants without residence (without cause: 51.1% vs 21.1% (χ 2 = 10.69, 1 df, p < .01); with cause: 71.3% vs 45.5% (χ 2 = 9.30, 1 df, p < .01)); and (3) the Spanish ethnic group indicated this option more than the Central/South Americans (without cause: 54.2% vs 25.4% (χ 2 = 17.89, 1 df, p < .001); with cause: 73.3% vs 49.3% (χ 2 = 14.47, 1 df, p < .001)). Using regression analysis, ethnicity (being Spanish) was a statistically significant predictor for advising the help of a priest for the sadness-without-cause scenario. On the other hand, conceptualizing sadness as a misfortune predicted seeking the help of a priest to face sadness with a cause. Interestingly, level of religious practice did not reach a statistically significant level for this prediction (Table 5).

Discussion

Distinction of normal sadness from dysfunctional sadness and help-seeking

The results confirmed our first hypothesis: most participants understood the character presenting with sadness without a perceived cause as being ill (and in particular mentally ill, along the lines of a depressive episode), while most of them understood the character undergoing sadness with a clear cause as facing a misfortune, as a normal response to the vicissitudes of life. In spite of both vignettes depicting characters undergoing similar symptoms (both fulfilling criteria for major depressive disorder), participants answered the question about what was happening to them in a very different manner, suggesting the key importance of having a cause triggering the symptoms.

The differences observed confirmed the distinction made by Horwitz and Wakefield (2007) between sadness with and without cause, as well as echoing Foster’s (1976) dichotomy of personalistic and naturalistic disease etiologies: personalistic causality involved an intervening agent (sadness with a clear cause, in this case, as a reaction to the daughter’s life-threatening condition), whereas naturalistic causality was theorized in impersonal terms (disease, endogenous depression).

Those conceptualizing sadness as an illness were more likely to recommend medical help-seeking (second hypothesis). Moreover, understanding sadness as an illness predicted seeking medical attention. It is interesting that understanding what the character was going through in pathological terms also predicted seeking help from family and friends, and not just from medical professionals. However, understanding the suffering of the character whose daughter is severely ill in non-pathological terms, but rather as a result of misfortune, predicted the suggestion that the character needed to ‘get better by themselves’. As the experience of sadness is not understood in pathological terms but as a normal reaction to adversity, the recommendation to find a way to deal with it by themself naturally follows.

Socio-cultural variation

The results revealed socio-cultural variation in both vignettes among age ranges, employment and education. Those who were in the oldest age range (above 65 years), had only school education and were unemployed, not only understood sadness more frequently as an illness, but also recommended help-seeking in tune with their conceptualization, having higher rates of suggesting the need for medical care. These differences might arise because those more educated, younger or employed were less prone to see the experience of sadness in a pathological way, as they might be more used to being self-reliant, having more of their own inner sources of help to resort to in times of sadness and also be in charge of those resources enabling them to deal with life problems in a non-pathological way. Availability of personal, intellectual, social or economic resources might make them less in need of adopting a sick role.

The experience of poverty, deprivation and marginalization may hinder the possibility of understanding sadness in a non-pathological way. An association between depression and poverty has been well established in research carried out in both industrialized and non-industrialized countries (Patel, 2001), revealing higher risk of depression in those who were unemployed (Bartley, 1994; Lewis & Sloggett, 1998), have lower incomes or standards of living (Chen et al., 2005; Lewis et al., 1998; Lorant et al., 2007; Weich & Lewis, 1998) and a lower level of education (Husain, Gater, Tomenson, & Creed, 2004; Mumford, Saeed, Ahmad, Latif, & Mubbashar, 1997).

Possible explanations regarding older people being more resistant to conceptualizing sadness outside an illness model could be due to frail health (with growing health problems reinforcing a more pathological view), isolation, reduced sense of purpose and less confidence in their own personal strength, among other reasons. Meta-analysis has shown that, among the elderly, poor self-reported health status is strongly associated with depression (Chang-Quan et al., 2010) and that bereavement, sleep disturbance and disability, among other variables, are important risk factors for depression (Cole & Dendukuri, 2003). A cross-sectional study among older people in rural China showed that depression was independently associated with lack of social support and poor health status (Chen et al., 2005).

It is interesting that those who were married or cohabiting were found to be more likely to see sadness with cause as a misfortune; furthermore, being married was also a predictor for conceptualizing sadness with cause as a misfortune. These findings reinforce the suggestion that the participants with existing sources of social support might be more likely to see the adoption of the sick role as less necessary, seeing the character as more able – maybe having had previous experience of counting on the help of their own spouse or partner – to understand the feelings of sadness in the face of adversity as a normal human reaction that can be overcome. On these lines, studies have found associations between depression and poor family relations (Chen et al., 2005) and with marital breakdown (Rotermann, 2007).

Spanish participants were more likely to advise recommending seeking support from relatives and friends than Central/South Americans in both scenarios. This difference could arise as the migrant participants might have left their families and friends behind in their home countries and were less ready to think about them as an immediate source of support for the characters.

Church engagement and help-seeking

Some studies have suggested that culturally mediated religious beliefs might deter people from accessing and engaging with mental health services, causing them to prefer spiritual care rather than a secular psychiatric alternative (Hatfield et al., 1996; McCabe & Priebe, 2004; Mitchell & Baker, 2000). However, this was not found in our study: recommending the help of a priest and medical attention reached very similar numbers (almost 50% for both sources of help) in the sadness-without-cause vignette. Moreover, those who never attended religious services or had no religious affiliation had approximately half the rate of recommending medical care for this vignette than those who attended church to some degree. This suggests that their church engagement at the very least does not hinder proposing help-seeking within the medical model and may even hint at some other process, for instance that religious people may seek help more frequently as they may have more hope in the possibility of recovery by some means or other. But this may also be explained in our study by the high level of education of the sample which, regardless of their religious views, might be more likely to give them a positive opinion of medical sciences.

On the other hand, those who attended church frequently suggested more often resorting to family and friends for dealing with sadness without cause than those who practised less or had no religious affiliation; the Catholic Church’s teachings, strongly emphasizing the central role of the family, may be behind this difference.

Help from the clergy

Our participants clearly saw their parish priests as legitimate providers of help when confronted with mental distress or illness. Those more religiously committed seemed to think that helping people deal with sadness is an integral part of their priest’s duty of pastoral care and would recommend relying on his help even in the secular experience of sadness (no information was given regarding the characters’ religious background and there was no religious content in their symptomatology). This raises complex questions regarding the boundaries of their pastoral care, the explanatory models for mental illness that priests may use and whether the training they received was adequate to enable them to recognize serious mental illnesses warranting psychiatric referral (Farrell & Goebert, 2008; Gilbert, 2007; Leavey, Loewenthal, & King, 2007).

Conceptualizing sadness as a misfortune was a predictor for advising the support of a priest: it is interesting that the priest is seen particularly as having a role in helping people coming to terms with life’s tragedies and challenges and as a central figure in their social support network. It is quite striking to consider that, for the sadness-with-cause scenario, seeking help from relatives/friends achieved virtually the same percentage (almost 70%) as seeking the support of a priest.

Statistically significant socio-cultural variations were found for recommending the help of a priest. Those participants whose ethnicity was Spanish and those whose legal status was defined as ‘Spanish nationals’ or ‘leave to remain’ chose this option more. A possible explanation for this variation may be that those who are migrants or without residence might have had more difficulties in integrating themselves in a religious community, owing to lack of time or lack of real or perceived acceptance from the communities in the host country. Interestingly, those living in local authority housing or who were homeless (living with friends or family) recommended seeking the advice of a priest more often; this finding may be attributed to the fact that the migrant centres included in the study are run by the church as charity organizations, so the participants may have had more regular access to the clergy and have seen them as useful. This may highlight the role of religious organizations filling a gap in services for more vulnerable members of society.

Those above 65 years old suggested seeking the help of the clergy more often than younger participants. This is understandable as they had the highest level of church engagement and it is in keeping with results of a recent national survey.[2]Although this survey showed that women were more regular churchgoers than men, no differences between the genders were found in our sample for seeking the help of a priest. This might be explained by the sample’s lack of gender differences for frequency of church attendance.

Limitations

There are, of course, limitations to this study. All participants came from Spain or other Spanish-speaking countries in Central or South America. It would have been interesting to have a more heterogeneous sample, particularly in terms of ethnicity and religious background (including other religions as well as more agnostic and atheist participants) to further examine cultural variation. On the other hand, the level of religious practice of the sample was not nationally representative as many were recruited from Catholic centres and were more religious than the general population. Nevertheless, a recent national survey looking at the frequency with which people went to mass or other religious services (excluding those occasions related to ceremonies of a social kind such as weddings or funerals) showed that over a sixth of the population had a high level of attendance (from several times a week to once weekly) and almost a quarter attended with some regularity (from at least one a month to several times a year) (Centre of Sociological Investigations, 2009a), therefore it may be generalizable to a significant proportion of the Spanish population. Moreover, as a measure of the participants’ religious observance, we relied on church attendance, which is a public aspect of one’s religion, not examining the role that the more personal and private aspects of one’s faith might have had. A further consideration is age as a variable: investigation of younger subjects may result in different results.

Hypothetical case vignettes were used in the survey. Although this is an established method for eliciting attitudes and beliefs, responses may not reliably describe their actual behaviour. On the other hand, presenting participants with the two scenarios in the same questionnaire may have facilitated the different answers obtained in the study as they were able to compare them (sadness with cause vs without cause). There was also a difference in the gender of the characters, although it is uncertain if or how this could have affected the results. In-depth interviews and data analysis using qualitative methods to investigate people’s experiences of sadness with and without cause, their conceptualization, coping strategies and resolution would have contributed to a more comprehensive understanding of the observed differences.

Conclusion

This survey has explored people’s understanding of sadness and help-seeking to further understand their views and the factors underpinning them. The failure of the current diagnostic classification to differentiate between the normal reaction to adversity and depressive disorder becomes apparent in this survey. If we were to apply diagnostic criteria to both characters, they would very likely receive a diagnosis of depressive disorder and be treated accordingly. But our participants clearly differentiated between both cases: the vignette depicting severe emotional distress precipitated by a loved one’s suffering was seen more frequently as a normal response, while the one portraying intense sadness in the absence of any cause was seen as pathological. Overall, the recommended help-seeking behaviours are also consistent with this distinction, with the medical option being more commonly recommended when there is no cause for sadness. This survey emphasizes the importance of taking into account the context in which the depressive symptoms occurred, as it seems that the absence of an appropriate context is what makes people conceptualize them as abnormal. It also raises questions regarding the lack of face validity of the current diagnostic classification for depressive disorder that exclusively uses descriptive criteria. Moreover, conceptualizing the character as ill was understandably found to be the main predictor for seeing the need for medical attention. Therefore, the current diagnostic criteria may hinder people’s normalization of sadness with cause, gradually changing their conceptualization of this to one where they see it as pathological. Furthermore, they may favour medical help-seeking behaviours (e.g. antidepressants) to the detriment of other personal and social resources that have been used to combat normal human sadness since ancient times, hindering their coming to terms with the cause of the sadness from an existential point of view.

This study has also shown that there is a socio-cultural variation in how people understand and deal with sadness; due to the particular composition of our sample, the important role that the clergy plays in assisting religious people undergoing deep sadness or depression has been revealed, highlighting the importance that the religious beliefs of patients have on their seeking of spiritual guidance during times of distress. Therefore, it is important that efforts are made by mental health professionals to carefully explore the patient’s own understanding of their symptoms and their wider social and cultural context, thereby gaining insight into the subjective experience of illness.