Mortality among immigrant patients 20–45 years of age with chronic back pain in primary care in Sweden: A 15-year follow-up cohort study

Abstract

BACKGROUND:

Mortality rates among immigrant patients undergoing rehabilitation for musculoskeletal backache are unknown.

OBJECTIVE:

To study the association between marital status, severe psychosocial strain, receiving long-term time-limited sickness allowance (TLSA) and all-cause mortality (ACM) in a cohort of immigrants aged 20–45 years with long-standing backache in Sweden.

METHODS:

We studied 318 patients (92% foreign-born, 76% non-European) of known marital status on sick-leave for musculoskeletal backache. They were followed up for ACM until 2015. Socio-demographic data, TLSA and psychosocial strain, including major depression, severe psychosocial stressors and pessimistic thoughts, were analysed using multiple-imputation Cox regression.

RESULTS:

Over a mean (standard deviation) follow-up time of 15 (5.0) years, 11 (3.5%) participants died. At baseline, 34% were unmarried, 19% were receiving TLSA, and 71% had $\geqslant$ 1 psychosocial strain component (38% depression; 47% severe stressors; 35% pessimistic thoughts). After concomitant risk factors were adjusted for, being unmarried and receiving TLSA were associated with higher mortality by factors of 6.2 ( $p=$ 0.005) and 5.8 ( $p=$ 0.006), respectively. Psychosocial strain was only significantly associated with higher mortality in the unadjusted analyses.

CONCLUSIONS:

Being unmarried and receiving TLSA were associated with significantly higher ACM in this highly marginalized group of immigrant patients.

Keywords

Chronic pain sickness allowance psychosocial stressors depression illness attitudes immigrants rehabilitation all-cause mortality primary care

1. Introduction

Long-standing musculoskeletal backache (in the lower or upper back or shoulder) is a multifactorial disorder with increasing co-morbidity with age [1]. Younger people may have other prominent negative factors associated with an elevated risk of poor outcomes of care and long-term sickness absence due to pain, such as ideas of life-threatening pain, poor health related to pain, other illness-related limitations and psychosocial stress, including acculturation stress among new immigrants [2, 3]. In a Swedish study, the overall mortality rate among younger middle-aged patients attributable to musculoskeletal and mental distress is not known, but mortality has increased among young people receiving disability pensions [4]. An increase in mortality has previously been shown among severely depressed men [5], individuals of low socioeconomic status [6], single men and drug users [7]. In Sweden, the rate of deaths attributed to musculoskeletal disorders as an underlying cause has decreased in recent years [8]. Patients in rehabilitation and on long-term sick-leave may suffer from covert morbidity and/or unhealthy lifestyles or social conditions, making it crucial to identify those in greater need of long-term support and medical attention in primary health care [9].

At the time of this study, time-limited sickness allowance (TLSA) [1] was granted to people in Sweden believed to be too sick for work because of pain in the subsequent 1 or 2 years. The underlying illness and disability were expected to be re-evaluated annually or biannually by a physician – often co-operating with a physiotherapist – to provide treatment, help the patient return to work and treat other health problem(s) [4]. In multicultural settings, this process should be adapted for patients from diverse socio-cultural and linguistic backgrounds [5, 6]. Multicultural city districts are often inhabited by many people requiring disability pensions owing to musculoskeletal pain and with short life expectancy over the age of 30 years, despite their often low alcohol consumption [7, 8].

Depression is a negative emotion often associated with higher mortality among psychiatric inpatients, but not among female outpatients [9]. Higher mortality has also been reported among younger people with lengthy periods of sickness absence [10] and those with low self-rated health [11], except for Swedish women [12]. Although the findings vary, immigration may also be a risk factor for death. However, in general, even new immigrants have similar illness patterns to natives [13]. Older people with self-rated limitations in daily functioning may have higher mortality rates [6, 14], although mortality has been found to be lower among ethnically diverse older people in distress or with widespread pain [15]. Low education is another risk factor for mortality, and is prevalent among many immigrant groups, especially women [8].

Most published studies are epidemiological. However, studies of mortality among immigrant patients in a clinical context are of interest for clinical workers aiming to improve health care in ethnically diverse settings. To the best of our knowledge, there are no studies of all-cause mortality (ACM) among younger middle-aged (below 45 years) ethnic and socio-culturally diverse primary care patients with long-standing backache, with or without comorbidity in the form of clinical depression, severe psychosocial stress and pessimistic attitudes towards pain, or those receiving TLSA, in relation to marital status.

1.1 Aims

We examined whether marital status, TLSA and severe psychosocial strain were independently associated with ACM among employed middle-aged immigrant patients treated in primary care for disabling long-standing backache.

1.2 Study hypothesis

We hypothesized that marital status, TLSA and severe psychosocial strain in the form of clinical depression, severe psychosocial stress and/or pessimistic thoughts about early death were independently associated with ACM.

2. Methods

A prospective cohort study was designed. No power calculation was performed because of the real-life design of the study. Baseline clinical data were derived from the first patient-doctor encounter in a 4-week rehabilitation programme conducted during 1994–2007 in a multicultural primary care setting (15,000 inhabitants: median age 29 years, 82% foreign-born from approximately 70 countries) in Stockholm, Sweden. For the full study, the baseline data were linked to the Swedish Population Register (PR) for follow-up of ACM throughout 2015. The Stockholm Regional Ethics Committee approved the full study (No. 2017/2345-32).

2.1 Baseline study: Summary

The target group included patients aged 20–45 years who had been on continuous sick-leave for $\geqslant$ 6 weeks for non-malignant pain disorders, including patients receiving TLSA. Individuals were accepted regardless of country of birth or Swedish language proficiency. Professional interpreters were available. The exclusion criteria were having potentially a deadly disease, being above the age limit, being on short-term sick-leave or being unemployed and not receiving financial compensation for sickness.

Physical, mental and psychosocial conditions, as well as attitudes to pain, were established during the initial assessments, which were followed by dialogue sessions about pain and daily physical training [2]. Daily or continuous work-hindering pain in the back and/or shoulders were the patients’ main health complaints. Some (6%) had co-morbidities such as mild anaemia, asthma or diabetes, with no significant differences between those receiving and not receiving TLSA. Paracetamol use, often in small doses, was common, whereas alcohol consumption was rare. Depression was diagnosed using diagnostic criteria from axis I in the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised (DSM-III-R), (DSM-III-R; 1987). Psychosocial stressors were scored according to the degree of severity, from 1 (lowest) to 6 (highest), and categorized as severe (score $>$ 2.5) or not severe (score $\leqslant$ 2.5) in accordance with axis IV in DSM-III-R. Attitudes to pain were established by standardized interviews, here categorized as pessimistic thoughts (yes/no) about imagined severe disability or the likelihood of early death from pain [16].

Researchers have preconceived notions or preconceptions about patients and results. This important issue was discussed by researchers in advance of the data sampling. Our preconceived notions were examined with curiosity and open minds to discover individual attitudes and conceptions of pain regardless of background. Patients were a mix of ordinary people from our daily practice, except for those with severe diseases. We conducted consecutive unselected sampling of patients who visited primary health care clinics for extended sick-leave certificates. Potential sources of bias were addressed using the standardized questionnaires and diagnostic methods that are common in rehabilitation programs. Additionally, liberal use of well-known and reliable interpreters was made to provide trustworthy translations.

2.2 Data collection

Data on age and gender were retrieved based on individual Swedish personal identification numbers (PINs), while dates of death or migration were obtained by linking the PINs to the PR maintained by the Swedish National Tax Board. The remaining data were obtained from personal interviews, medical examinations, medical records, the National Social Insurance Agency and personal accounts of pain and function. During the study, 456 patients were included in consecutive order and gave consent to participate; another 28 refused to participate. This study focused on the 341 patients in the rehabilitation programme with valid PINs, for whom it was possible to obtain data relating to ACM from the PR. A flowchart of the inclusion process is shown in Fig. 1.

Figure 1.

Flowchart for inclusion of participants in the study.

2.3 Outcomes

Time-to-event (TTE) was the main outcome in the current study, with the event of interest being ACM and the date of follow-up set to December 9, 2015. TTE was calculated as the time from the date of inclusion in the study to the date of death, emigration or follow-up. Individuals who emigrated from Sweden before follow-up or who were still alive at the date of follow-up were considered to be censored at the day of emigration or follow-up, respectively.

2.4 Risk factors

The risk factors of main interest in the current study were receiving TLSA (yes/no) and level of psychosocial strain, with the latter calculated based on the three variables of depression (yes/no), severe psychosocial stressors (yes/no) and pessimistic thoughts (yes/no). For each of these three variables, the patient was given 1 point for each ‘yes’ and 0 points for each ‘no’ response, resulting in a total of 0–3 points for psychosocial strain. Secondary risk factors, included as control variables in this study, were age (years), gender (male/female), marital status (married/not married), region of origin (Africa, Asia, Northern Europe, Southern Europe, Middle East or South America), use of an interpreter (yes/no), number of children at home, number of years of schooling and confidence in a cure (yes/no). In the categorization of patients’ regions of origin, Northern Europe included Belgium, Finland, France, Poland, Sweden and Switzerland, while Southern Europe included Greece, Italy and the former Yugoslavia. Individuals from Iraq, Iran, Jordan, Lebanon or Syria were categorized as being from the Middle Eastern region, those from Afghanistan, Bangladesh, India, Pakistan or Turkey as being from Asia and those from Algeria, Eritrea, Ethiopia, Morocco, Somalia, Tunisia or Uganda as being from Africa.

2.5 Statistical analyses

Categorical data are presented as frequencies and percentages, $n$ (%), while discrete and continuous data are given as means with accompanying standard deviations (SDs). Tests for differences between married and unmarried were performed using Pearson’s $\chi^{2}$ test for categorical data, the Mann-Whitney $U$ -test for discrete and continuous data and the log-rank test for TTE data. For these analyses, only individuals with valid data for the particular variable in question were used in the analyses. Adjusted and unadjusted multiple imputation Cox regression models with 50 imputations were used to examine the associations between the risk factors and the outcome TTE. The results are presented as a pooled hazard ratios with accompanying 95% confidence intervals. Unadjusted models were calculated for all risk factors, and for comparison, were calculated separately for the three components of the psychosocial strain index. All risk factors with $p$ values $<$ 0.20 (excluding the separate components of the psychosocial strain index) were simultaneously included in an adjusted Cox regression model, and the model was trimmed using a backward elimination procedure whereby all risk factors with $p$ values $\geqslant$ 0.20 were excluded one at a time. In the regression analyses, region of origin was dichotomized as within or outside Europe. All statistical analyses were performed using IBM SPSS Statistics (ver. 23/24; IBM Corp., Armonk, NY, USA), with two-sided $p$ values $<$ 0.05 considered statistically significant.

3. Results

Table 1 shows the basic characteristics of the study population. The mortality rate was low, with 11 (3.5%) deaths during a mean (SD) follow-up period of 14.9 (5.0) years, perhaps attributable to the youth of the participants, who had a mean (SD) age of 38.2 (6.6) years. Most of the participants were women, $n=$ 204 (64.2%), with Turkish immigrants being the largest minority, $n=$ 91 (28.6%). Only 25 (7.9%) participants reported that Sweden was their country of origin. Nearly all worked in services, cleaning and care-giving occupations (not listed in Table 1).

Table 1
Basic characteristics of the study population ( $n=$ 318)

Variable	Total		Married		Unmarried		$p$ value ${}^{\text{a}}$		Missing
	( $n=$ 318)		( $n=$ 210)		( $n=$ 108)				$n$ (%)
Status at follow-up, $n$ (%)								0.018 ${}^{\text{b}}$	0	(0.0)
Living	291	(91.5)	198	(94.3)	93	(86.1)
Emigrated	16	(5.0)	8	(3.8)	8	(7.4)
Deceased	11	(3.5)	4	(1.9)	7	(6.5)
Age (y), mean (SD)	38.2	(6.6)	38.2	(6.2)	38.2	(7.2)		0.768	0	(0.0)
Female, $n$ (%)	204	(64.2)	137	(65.2)	67	(62.0)		0.573	0	(0.0)
Non-European	243	(76.4)	164	(78.1)	79	(73.1)		0.325	0	(0.0)
Region of origin, $n$ (%)							$<$	0.001	0	(0.0)
Africa	45	(14.2)	20	(9.5)	25	(23.1)
Asia	111	(34.9) ${}^{\text{c}}$	91	(43.3)	20	(18.5)
Middle East	73	(23.0)	50	(23.8)	23	(21.3)
Northern Europe	35	(11.0) ${}^{\text{d}}$	15	(7.1)	20	(18.5)
Southern America	14	(4.4)	3	(1.4)	11	(10.2)
Southern Europe	40	(12.6)	31	(14.8)	9	(8.3)
Interpreter, $n$ (%)	86	(29.6)	63	(32.8)	23	(23.2)		0.090	27	(8.5)
Number of children at home, mean (SD)	2.4	(1.7)	2.7	(1.6)	1.9	(1.6)	$<$	0.001	15	(4.7)
Schooling (y), mean (SD)	8.3	(4.3)	7.7	(4.4)	9.5	(3.9)	$<$	0.001	25	(7.9)
Confidence in a cure, $n$ (%)	196	(72.6)	133	(74.7)	63	(68.5)		0.276	48	(15.1)
Time-limited sickness allowance, $n$ (%)	61	(19.2)	47	(22.5)	14	(13.0)		0.041	1	(0.3)
Depression, $n$ (%)	115	(37.7)	75	(37.5)	40	(38.1)		0.919	13	(4.1)
Severe psychosocial stressors, $n$ (%)	135	(47.0)	75	(39.7)	60	(61.2)		0.001	31	(9.7)
Pessimistic thoughts, $n$ (%)	98	(34.6)	64	(34.8)	34	(34.3)		0.941	35	(11.0)
Psychosocial strain (0–3 points) ${}^{\text{e}}$								0.061	46	(14.5)
0 points, $n$ (%)	78	(28.7)	56	(31.6)	22	(23.2)
1 point, $n$ (%)	98	(36.0)	60	(33.9)	38	(40.0)
2 points, $n$ (%)	68	(25.0)	48	(27.1)	20	(21.1)
3 points, $n$ (%)	28	(10.3)	13	(7.3)	15	(15.8)

${}^{\text{a}}P$ values $<$ 0.05 are given in bold. $P$ values for differences between married and unmarried were calculated using Pearson’s $\chi^{2}$ test for categorical data, the Mann-Whitney $U$ -test for discrete and continuous data and the log-rank test for time-to-event data. ${}^{\text{b}}$ Deceased vs living or emigrated. ${}^{\text{c}}$ Of whom, 91 (28.6%) were born in Turkey. ${}^{\text{d}}$ Of whom, 25 (7.9%) were born in Sweden. ${}^{\text{e}}$ One point for each ‘yes’ for severe psychosocial stressors, pessimistic thoughts and being depressed.

Regarding the main risk factors, one-third ( $n=$ 108; 34.0%) of the participants were unmarried, and about one-fifth ( $n=$ 61; 19.2%) were receiving TLSA at the start of the study. About half of the patients, $n=$ 135 (47.0%), suffered from severe psychosocial stressors, including anxiety relating to pain. In total, 98 (34.6%) patients had pessimistic thoughts, while 115 (37.7%) were classified as clinically depressed. Very few abused drugs, and only a few had a blood pressure reading $>$ 130/80 mmHg (not listed in Table 1).

ACM was significantly ( $p=$ 0.018) higher among unmarried (6.5%) than among married (1.9%) participants. There were more unmarried individuals from Africa, Northern Europe and Southern America, and fewer from Asia and Southern Europe. Compared with their married counterparts, the unmarried participants had significantly fewer children living at home ( $p<$ 0.001), more years in school ( $p<$ 0.001) and were less often receiving TLSA ( $p=$ 0.041), but were more likely to suffer from severe psychosocial stressors ( $p=$ 0.001).

The results from the Cox regression analyses are shown in Table 2. In the unadjusted analyses, a significantly increased risk of ACM during the follow-up period was observed for unmarried individuals ( $p=$ 0.028) and those receiving TLSA ( $p=$ 0.016), with 4.0- and 4.4-times higher risks, respectively. None of the separate components of the psychosocial strain index – clinical depression ( $p=$ 0.285), pessimistic thoughts ( $p=$ 0.272) and severe psychosocial stressors ( $p=$ 0.058) – were significantly associated with increased ACM. However, the psychosocial strain index was associated with significantly increased ACM, with twice the risk of death for participants during the follow-up period for each additional component ( $p=$ 0.034). None of the secondary risk factors was significantly associated with ACM during the follow-up period.

Table 2

Results of Cox regression analysis for predicting time-to-event showing pooled unadjusted and adjusted hazard ratios (HRs) with accompanying 95% confidence intervals (CIs) and $p$ values for multiple imputation models using 50 imputations for missing values

Variable	HR (95% CI)		$p$ value	HR (95% CI)	$p$ value
	Unadjusted			Adjusted
Age (y) ${}^{\text{a}}$	1.039	(0.942–1.145)	0.445	_ ${}^{\text{b}}$
Female ${}^{\text{a}}$	2.351	(0.508–10.886)	0.274	_ ${}^{\text{b}}$
Unmarried ${}^{\text{a}}$	3.959	(1.157–13.543)	0.028	6.199 (1.743–22.049)	0.005
Non-European ${}^{\text{a}}$	0.392	(0.119–1.286)	0.122	0.285 (0.077–1.055)	0.060
Interpreter	2.480	(0.661–9.311)	0.178	_ ${}^{\text{b}}$
Number of children at home	0.923	(0.598–1.424)	0.715	_ ${}^{\text{b}}$
Schooling (y)	0.900	(0.773–1.048)	0.175	0.861 (0.720–1.030)	0.102
Confidence in a cure	1.005	(0.161–6.266)	0.996	_ ${}^{\text{b}}$
Time-limited sickness allowance	4.394	(1.314–14.696)	0.016	5.773 (1.656–20.122)	0.006
Psychosocial strain (0–3 points) ${}^{\text{c}}$	1.989	(1.054–3.755)	0.034	_ ${}^{\text{b}}$
Depression	1.928	(0.579–6.415)	0.285	_ ${}^{\text{b}}$
Pessimistic thoughts	2.041	(0.571–7.299)	0.272	_ ${}^{\text{b}}$
Severe psychosocial stressors	4.544	(0.951–21.711)	0.058	_ ${}^{\text{b}}$

Note: Results based on 11 events and 305 censored participants (two individuals were censored before the earliest event in a stratum). $P$ values $<$ 0.05 are given in bold. ${}^{\text{a}}$ No missing values. ${}^{\text{b}}$ Not in model. ${}^{\text{c}}$ One point for each ‘yes’ for severe psychosocial stressors, pessimistic thoughts and being depressed.

In the adjusted analyses, being unmarried and receiving TLSA showed independent significant associations with an increased risk of ACM: the risk was 6.2 times higher for unmarried compared with married participants ( $p=$ 0.005), and 5.8 times higher for those receiving compared with those not receiving TLSA ( $p=$ 0.006). However, psychosocial strain did not retain its significant association and did not make it to the final adjusted model. None of the secondary risk factors were significantly associated with ACM.

4. Discussion

The present long-term cohort study of younger middle-aged and primarily immigrant patients with long-standing non-malignant pain found that being unmarried or receiving TLSA when entering the study was associated with a significantly increased risk of ACM during the 15-year follow-up period, after adjusting for relevant secondary risk factors. Psychosocial strain, in the form of clinical depression, severe psychosocial stress and/or pessimistic thoughts about future life, was only significantly associated with increased mortality in the unadjusted analyses.

4.1 Discussion of results

There were few deaths in this socio-culturally diverse patient cohort, which may be attributed to the young age of the patients (mean age 38.2 years at inclusion) and the low prevalence of alcohol consumption, drug abuse and severe co-morbidities. Thus, this cohort may be considered to consist of ‘healthy workers’. However, being unmarried (about one-third of the participants) was a predictor of higher mortality rates, despite having higher levels of education, fewer children living at home, less TLSA but more severe psychosocial stressors. Notably, TLSA and unmarried status were the main predictors of death. This is a new focus of research in sickness-related work absence, whereby receiving TLSA seems to be a major predictor of future death, while disabling musculoskeletal pain alone does not seem to be strongly associated with increased risk of mortality [17].

An innovative feature of the present study was the use of a multicultural patient sample and of observed data, and clinically systematically (as opposed to self-) rated, personal risk factors such as serious marital, financial or personal problems and trauma, which constitute a sum of the psychosocial stressors that are building blocks of the complex pattern of long-standing pain. There are also good news, that is that the observed low mortality in this sample could be attributed to two well-known factors mitigating the risk of early death: a low prevalence of alcohol consumption and a high marriage rate.

A trend toward increased mortality was seen in our data when we combined the explanatory variables of depression, pessimistic thoughts and severe psychosocial stress into an index of psychosocial strain. This weak association might have been stronger if there had been more cases of patients with severe depression. There were only two deaths among the patients assessed to have been exposed to moderate or few stressors. Thus, the association between severe stress or depression and mortality may have been under-estimated. Furthermore, for some patients, treatment after the baseline period may have alleviated their pessimistic thoughts and depressive mood [16]. A longer follow-up period with more events may also have provided increased statistical power, rendering the results significant even in adjusted analyses.

4.2 Discussion of methods

Because most participants in this study belong to highly marginalized groups in Swedish society, there were some inherent problems with the reliability of the data on mortality and emigration. A major problem of the Swedish PR is that some data regarding emigration and death outside Sweden rely on voluntary reporting, meaning that it is entirely possible to leave Sweden, die in another country and still be registered in the PR as alive and living in Sweden. The over-coverage in the PR (i.e. an individual being registered as living in Sweden despite having moved abroad) of individuals born outside the Nordic countries has been estimated to be 8%, resulting in low observed mortality figures for foreign-born individuals in the PR database [18]. Given that most participants in our study were immigrants, the frequency of only 11 (3.5%) deceased people may have been somewhat higher. A longer follow-up period with more events may also have resulted in increased power, rendering the results significant even in adjusted analyses. On the other hand, according to statistics from Stockholm County, the multicultural districts in western Stockholm with a high proportion of individuals from immigrant backgrounds have life expectancies of both women and men similar to that of the country in general [14].

4.3 Contribution to the literature

Our study contributes to knowledge regarding ACM rates among younger or middle-aged adults of mixed socio-cultural backgrounds with long-standing pain, with a focus on people under psychosocial strain receiving TLSA. Our study also highlights some inherent problems with the reliability of the data on mortality and emigration in Sweden.

4.4 Limitations and strengths

A major limitation of the study was the low number of observed events that, together with the small sample size, resulted in low statistical power. The study was also hampered by missing clinical and follow-up data. Our study sample had only a small number of second-generation immigrants, making it less feasible to run separate analyses for first- and second-generation immigrants. Future studies should investigate the effects of acculturation and whether differences exist between such groups in the context of back pain. Furthermore, our study used consecutive sampling to recruit participants, and included only employed or self-employed participants, making it unrepresentative of the population at large.

Among the strengths of our study were the comprehensive diagnostic data, the long period to follow-up (up to 15 years) and the frequent use of professional interpreters, which helped to enhance the validity of the interview data. A further strength was the many recent first-generation immigrants in our sample with special problems in meeting the demands of the socio-cultural environment of a new country.

Our results should be interpreted with caution, not least because of the high numbers of participants excluded from follow-up because of invalid PINs, thus biasing the study sample. However, we believe that the present results could be applicable to similar multicultural patient groups in primary health care in other countries, and may help to both identify patient subgroups at risk of increased mortality and inform future projects. In addition, our results may currently be valid for rehabilitation care, considering the increasing migrant populations of segregated areas in big cities [8]. Moreover, in Sweden, mental and musculoskeletal disorders remain major reasons for sick-leave and create difficulties for doctors seeking to evaluate, treat and predict the duration and severity of illness and verify patients’ inability to work [19].

4.5 Comparison with other studies

Being male, never marrying or having illness-related limitations are risk factors for death, in addition to age [20], but we found no gender difference in mortality. Many immigrant groups living in Europe, especially women, seem to have lower mortality rates than the native population [21], except for immigrants from sub-Saharan countries [22], although such people were rare in our study cohort. Socio-economic differences may account for most group and ethnic differences [23, 24, 25] in Nordic countries [6, 26]. Studies of chronic pain and mortality often lack consistent definitions of chronic pain, social data and psychological factors, making it difficult to carry out comparisons [27]. However, in general, chronic and widespread pain does not seem to be associated with increased mortality [28, 29], except when combined with low self-rated health and adverse lifestyles [30]. Then again, an earlier review of chronic pain and mortality found mildly increased mortality rates from cancer in some studies [27], and more deaths among middle-aged people with pain in the chest or abdomen [28]. Notably, the participants in the present study very seldom reported chest or abdominal pain.

Depression in older men increases their risk of death [31]. Another risk factor for death is physical pain associated with thoughts of escape from life [32, 33] or hopelessness, misuse of drugs and additional medical conditions [32, 34, 35]. Moreover, middle-aged primary care patients with depression are at an especially high risk of death from unnatural or alcohol-related causes [36]. Pessimism regarding future health can also be a risk factor for death among young men [37]. By comparison, 4.5% of our cohort with combined physical pain and depression had died.

4.6 Clinical implications

We should be aware that younger middle-aged patients with back pain may also face serious future complications, especially those who are unmarried, receiving TLSA or have components of severe psychosocial strain. These individuals may benefit from regular medical follow-up consultations after rehabilitation.

4.7 Future studies

Repeated studies of primary care patients consisting of larger samples with precautions against loss of opportunity to follow-up and the inclusion of more clinical data are needed to confirm our results.

5. Conclusions

The present study found that among predominantly younger middle-aged immigrant patients, being unmarried and receiving long-term TLSA were associated with a significantly increased risk of ACM in adjusted analyses, whereas psychosocial strain was associated with increased mortality only in unadjusted analyses.

Footnotes

Acknowledgments

Approval for the baseline and follow-up studies was obtained from the North Stockholm Ethics Committee (Dnr 00-166) and the Stockholm Regional Ethics Committee (Dnr 2017/2345-32), respectively. The study was funded by departmental resources.

Conflict of interest

None to report.

References

Swedish Council on Technology Assessment in Health Care. Sickness absence: Causes, consequences, and physicians’ sickness certification practice. A systematic literature review by the Swedish Council on Technology Assessment in Health Care (SBU). Scand J Public Health Suppl. 2004; 63: 3–263. PubMed PMID: 15513647.

Löfvander

Engström

. “Unable and useless” or “able and useful”? A before-and-after study in the primary care of self-rated inability to work in young immigrants having long-standing pain. J Back Musculoskelet Rehabil. 2004; 17: 91–100.

Mantwill

Schulz

. Causal attribution and coping maxims differences between immigrants and non-immigrants suffering from back pain in Switzerland. PLoS One. 2016; 11: e0161758. PubMed PMID: 27583445. PMCID: PMC50087 24 n.

Ljungquist

Hinas

Arrelöv

Lindholm

Wilteus

Nilsson

, et al. Sickness certification of patients – a work environment problem among physicians? Occup Med (London). 2013; 63: 23–9.

Taloyan

Al-Windi

Johansson

Saleh-Stattin

. Kurdish men’s experiences of migration-related mental health issues. Prim Health Care Res Dev. 2011; 12: 335–47. PubMed PMID: 22284948.

Pottie

. Misinterpretation: Language proficiency, recent immigrants, and global health disparities. Can Fam Physician. 2007; 53: 1899–901. PubMed PMID: 18000266. PMCID: PMC2231483.

Johansson

Helgesson

Lundberg

Nordquist

Leijon

Lindberg

, et al. Work and health among immigrants and native Swedes 1990–2008: A register-based study on hospitalization for common potentially work-related disorders, disability pension and mortality. BMC Public Health. 2012; 12: 845. PubMed PMID: 23039821. PMCID: PMC3532317.

Life circumstances, life habits and health in county of Stockholm län (Levnadsvillkor, levnadsvanor och hälsa i Stockholms län). Solna: Centrum för epidemiologi och samhällsmedicin. Rapport 2014: 3. www.fokhalsoguiden.se. In Swedish.

Wedegaertner

Arnhold-Kerri

Sittaro

Bleich

Geyer

Lee

. Depression- and anxiety-related sick leave and the risk of permanent disability and mortality in the working population in Germany: A cohort study. BMC Public Health. 2013; 13: 145.

10.

Mittendorfer-Rutz

Kjeldgård

Runeson

Perski

Melchior

Head

, et al. Sickness absence due to specific mental diagnoses and all-cause and cause-specific mortality: A cohort study of 4.9 million inhabitants of Sweden. PLoS One. 2012; 7: e45788.

11.

Halford

Wallman

Welin

Rosengren

Bardel

Johansson

, et al. Effects of self-rated health on sick leave, disability pension, hospital admissions and mortality. A population-based longitudinal study of nearly 15,000 observations among Swedish women and men. BMC Public Health. 2012; 12: 1103. doi: 10.1186/1471-2458-12-1103.

12.

Hange

Mehlig

Lissner

Guo

Bengtsson

Skoog

, et al. Perceived mental stress in women associated with psychosomatic symptoms, but not mortality: Observations from the Population Study of Women in Gothenburg, Sweden. Int J Gen Med. 2013; 6: 307–15.

13.

Halldorsdottir

Jonsson

Gudmundson

. A few observations on health service for immigrants at a primary health care centre. Int J Family Med. 2016; 2016: 6963835. doi: 10.1155/2016/6963835.

14.

Dauvrin

Lorant

. Adaptation of health care for migrants: Whose responsibility? BMC Health Serv Res. 2014; 14: 294. PubMed PMID: 25005021. PMCID: PMC4108228.

15.

Gurland

Teresi

Eimicke

Maurer

Reid

. Quality of life impacts on 16-year survival of an older ethnically diverse cohort. Int J Geriatr Psychiatry. 2014; 29: 533–45. PubMed PMID: 24167085. PMCID: PMC4372056.

16.

Löfvander

Engström

Theander

Furhoff

. Rehabilitation of young immigrants in primary care. A comparison between two treatment models. Scand J Prim Health Care. 1997; 15: 123–8. PubMed PMID: 9323778.

17.

Gjesdal

Ringdal

Haug

Maeland

Vollset

Alexanderson

. Mortality after long-term sickness absence: Prospective cohort study. Eur J Public Health. 2008; 18: 517–21.

18.

Ludvigsson

Almqvist

Bonamy

Ljung

Michaelsson

Neovius

, et al. Registers of the Swedish total population and their use in medical research. Eur J Epidemiol. 2016; 31: 125–36. PubMed PMID: 26769609.

19.

Carlsson

Lännerström

Wallman

Holmström

. General practitioners’ perceptions of working with the certification of sickness absences following changes in the Swedish social security system: A qualitative focus-group study. BMC Fam Pract. 2015; 16: 21. PMCID: PMC4339246 doi: 10.1186/s12875-015-0238-5.

20.

Reile

Stickley

Leinsalu

. Large variation in predictors of mortality by levels of self-rated health: Results from an 18-year follow-up study. Public Health. 2017; 145: 59–66.

21.

Helgesson

Johansson

Nordqvist

Lundberg

Vingard

. Sickness absence at a young age and later sickness absence, disability pension, death, unemployment and income in native Swedes and immigrants. Eur J Public Health. 2015; 25(4): 688–92. PubMed PMID: 25634955. PMCID: PMC4512957.

22.

Fedeli

Ferroni

Pigato

Avossa

Saugo

. Causes of mortality across different immigrant groups in Northeastern Italy. Peer J. 2015; 3: e975.

23.

Klinthäll

Lindström

. Migration and health: A study of effects of early life experiences and current socio-economic situation on mortality of immigrants in Sweden. Ethn Health. 2011; 16(6): 601–23. PubMed PMID: 21806407.

24.

Campbell

Edwards

. Ethnic differences in pain and pain management. Pain Manag. 2012; 2(3): 219–30. PubMed PMID: 23687518. PMCID: PMC3654683.

25.

Hollander

Bruce

Ekberg

Burström

Borrell

Ekblad

. Longitudinal study of mortality among refugees in Sweden. Int J Epidemiol. 2012; 41(4): 1153–61. PubMed PMID: 22552872.

26.

Norredam

Olsbjerg

Petersen

Juel

Krasnik

. Inequalities in mortality among refugees and immigrants compared to native Danes – a historical prospective cohort study. BMC Public Health. 2012; 12: 757.

27.

Smith

Wilkie

Uthman

Jordan

McBeth

. Chronic pain and mortality: A systematic review. PLoS One. 2014; 9(6): e99048.

28.

Kåreholt

Brattberg

. Pain and mortality risk among elderly persons in Sweden. Pain. 1998; 77(3): 271–8.

29.

Åsberg

Heuch

Hagen

. The mortality associated with chronic widespread musculoskeletal complaints: A systematic review of the literature. Musculoskeletal Care. 2017; 15(2): 104–13.

30.

Case

Deaton

. Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century. Proc Natl Acad Sci U S A. 2015; 112(49): 15078–83. PubMed PMID: 26575631. PMCID: PMC4679063.

31.

Holwerda

van Tilburg

Deeg

Schutter

Van

Dekker

, et al. Impact of loneliness and depression on mortality: Results from the longitudinal ageing study amsterdam. Br J Psychiatry. 2016; 209(2): 127–34. PubMed PMID: 27103680.

32.

Hooley

Franklin

Nock

. Chronic pain and suicide: Understanding the association. Curr Pain Headache Rep. 2014; 18(8): 435. PubMed PMID: 24916035.

33.

Tang

Crane

. Suicidality in chronic pain: A review of the prevalence, risk factors and psychological links. Psychol Med. 2006; 36(5): 575–86. PubMed PMID: 16420727.

34.

Vilhjalmsson

Kristjansdottir

Sveinbjarnardottir

. Factors associated with suicide ideation in adults. Soc Psychiatry Psychiatr Epidemiol. 1998; 33(3): 97–103.

35.

Aagaard

Buus

Wernlund

Foldager

Merinder

. Clinically useful predictors for premature mortality among psychiatric patients visiting a psychiatric emergency room. Int J Soc Psychiatry. 2016; 62(5): 462–70. PubMed PMID: 27095414.

36.

Strömberg

Backlund

Johansson

Löfvander

. Mortality in depressed and non-depressed primary care Swedish patients: A 12-year follow-up cohort study. Fam Pract. 2013; 30(5): 514–9. PubMed PMID: 23913789.

37.

Larsson

Hemmingsson

Allebeck

Lundberg

. Self-rated health and mortality among young men: What is the relation and how may it be explained? Scand J Public Health. 2002; 30(4): 259–66. PubMed PMID: 12680501.

Mortality among immigrant patients 20–45 years of age with chronic back pain in primary care in Sweden: A 15-year follow-up cohort study

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

1.1 Aims

1.2 Study hypothesis

2. Methods

2.1 Baseline study: Summary

2.2 Data collection

2.4 Risk factors

2.5 Statistical analyses

3. Results

Table 1 Basic characteristics of the study population ( n = 318)

4.1 Discussion of results

4.2 Discussion of methods

4.3 Contribution to the literature

4.4 Limitations and strengths

4.5 Comparison with other studies

4.6 Clinical implications

4.7 Future studies

5. Conclusions

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Basic characteristics of the study population ( $n=$ 318)