Round 1

Reviewer 1 Report

The topic of research is highly relevant. The methodology as described suggests solid research. The sample seems somewhat small and dispersed: 250 families, of which 200 are immigrant families / 50 native. Divided between two major cities, 100 / 25 families per city over the time span of 3 years (2015 - 2017). 

Some questions and suggestions per line: 

34-35: same year is mentioned, 1998

35: is the relationship between the global financial crisis and the end of growth so direct? perhaps other factors could be mentioned? 

42: to move "The 2008-post-crisis.." to an earlier part (around line 39)?

44: the quality of work - should it be working conditions? 

46: what does "low social support" entail? same for "high job insecurity" and "high psychological demands"   

47: "poor quality jobs" - such as?

49: "work from Spain" - research findings / research conducted in Spain? 

52: "other work" - ?

52: sentence unclear

55: "a second" - what? what is first? 

57: "another" - ? 

60: spelling mistake 

96: "the outcome was common mental disorders" - ?

204-205: "other work" - research findings? 

221: "these findings question the acculturation theory" - should it be mentioned in the introduction that this research paper works towards contesting the theory? 

Conclusion: to admit the limitations of the conclusion, given the size of the sample? 

Author Response

Response to Reviewer 1 Comments

Point 1: The topic of research is highly relevant. The methodology as described suggests solid research. The sample seems somewhat small and dispersed: 250 families, of which 200 are immigrant families / 50 native. Divided between two major cities, 100 / 25 families per city over the time span of 3 years (2015 - 2017). 


Response 1: We have included the following as a new paragraph in the material and method section to explain how we estimated the sample size (L93).

Sample size was estimated with a Poisson approximation. Three hundred and fifty nine immigrants and 107 natives were required to detect a minimum relative risk of 1.5 (based on earlier work), assuming an alpha of 0.05 and beta of 0.2 and an attrition rate of 15%. Assuming four family members per family (an estimated two children per family, based on the number of children per woman of foreign origin in Spain 1,54 [2]) and rounding upwards gives 180 immigrant and 50 Spanish families in each city.

Point 2: 34-35: same year is mentioned, 1998

Response 2. Thanks for pointing this out – it has been changed to 1988

Point 3: 35: is the relationship between the global financial crisis and the end of growth so direct? perhaps other factors could be mentioned?

Response 3: Yes, we have added a reference to back up that statement.  

Point 4: 42: to move "The 2008-post-crisis.." to an earlier part (around line 39)?

Response 4: Apologies, this sentence was not clear. We have left it in its original position but amended it to read as follows.

However, the post-crisis situation has resulted in a highly precariat labor market affecting the entire salaried population. 

Point 5: 44: the quality of work - should it be working conditions?

Response 5: Quality of work refers to working conditions and employment arrangements.

We have modified this sentence to read as follows.

“The quality of work (both in term of working conditions and employment arrangements) is an important determinant of mental health”

Point 6: 46: what does "low social support" entail? same for "high job insecurity" and "high psychological demands"  

Response 6: It has been clarified as follows:

Adverse working conditions associated with poorer mental health include job strain (high demand and low control), low social support (low social interaction in the workplace with either colleagues or supervisors)[5], high job insecurity (high probability of losing your job in the near future) [6] and high psychological demands(e.g. job strain (high demand/low control) [7].

Point 7: 47: "poor quality jobs" - such as?

Response 7: This sentence has been amended to read.

Point 8: 49: "work from Spain" - research findings / research conducted in Spain?

Response 8: Changed

Point 9: 56: "other work" - ?

Response 9 Sentence amended to other research

Point 10: 52: sentence unclear

Response 10: This sentence has been amended to

Work for migrant workers is particularly important as the chance to improve their economic conditions is the main driver for migration [9].

Point 11: 55: "a second" - what? what is first?

Response 11: The first explanation is acculturation.

Point 12: 57: "another" - ?

Response 12: A second explanation

Point 13: 60: spelling mistake

Response 13: Unclear what word you think has been misspelled?

Point 14: 96: "the outcome was common mental disorders" - ?

Response 14: This sentence was removed and this section now starts with

Common mental disorders were measured at the baseline 2015 and one year later in 2016 using the Spanish-language validated version of the 12-item General Health Questionnaire (GHQ-12) with four response options [16].

Point 15: 204-205: "other work" - research findings?

Response 15: changed to other research

NB we use the term other work in order to not constantly repeat the terms other research or other studies.

Point 16: 221: "these findings question the acculturation theory" - should it be mentioned in the introduction that this research paper works towards contesting the theory?

Response 16: In the final paragraph of our introduction we list our objectives which include ‘taking into account the duration of residence’. It was not our original intention contest the theory.

Point 17: Conclusion: to admit the limitations of the conclusion, given the size of the sample?

Response 17: A sentence in the conclusion has been amended to read

Bearing in mind that these findings come from a study with a small number of participants policies aimed at preserving the mental health of new arrivals and improving the mental health of Spanish-born and longer duration of residence immigrants are warranted.

Author Response File: Author Response.docx

Reviewer 2 Report

This study aims to investigate the association between new episodes of mental disorders and the duration of residence among migrant workers in Spain. Given the scale of the problem and the resumed migration of workers after the financial crisis of 2008, this is an important topic to investigate. Using the data from the PELFI study, the authors followed-up 28 Spanish-born and 102 immigrant workers who reported good mental health at baseline for 1 year. Incidence of common mental disorders was the highest in the Spanish-born workers. Immigrants residing in Spain for more than 15 years showed a similar incidence rate as the Spanish-born workers and the incidence of common mental disorders decreased with less residence time in Spain. This information can influence policy-makers to invest more resources in mental health prevention programs.  Although it is a valuable paper, a few changes are necessary to make it acceptable for publication. Comments below are intended to further strengthen the paper.

Abstract:

The authors could consider adding a sentence summarizing their conclusions to the abstract.

Introduction:

L34 – There is probably an error in one of the “1998’s”

L41-42 – “and 1.1768.36 worker workers have a work contract (16.2% of all the contracts)[1].” – not clear

The authors should explain why they focused on this particular cohort (Colombian and Ecuadorian). Also, earlier in the text they provided 3 alternative hypotheses for the decline of migrant health. Do the authors plan to test any of them in this study?

Methods:

Operationalization of “good/poor mental health”. Why was the cut-off score set at 3? Was it based on the prior literature?

Only 130 participants (of the 473 interviewed at baseline) were employed and in good mental health at baseline. That is only 27% of the sample. Was this mostly driven by the employment status or by the poor health of the respondents? Could the authors comment on the high prevalence of poor health in their sample and compare it with what was reported in the literature?

L106: typo “34”

Why did the authors decide on the following categories: 1-10, 11-15, and above 15? How many respondents were in each of the categories?

Discussion:

L195 – LA: abbreviation not explained

L207 - It in not necessary to spell out “confidence interval” at this point

Author Response

Response to Reviewer 2 Comments

Point 1: Abstract: The authors could consider adding a sentence summarizing their conclusions to the abstract.

Response 1: We think the last sentence of the abstract summarises our conclusions effectively.

Point 2: Introduction: L34 – There is probably an error in one of the “1998’s”

 Response 2: The second has been changed to 1988.

Point 3: L41-42 – “and 1.1768.36 worker workers have a work contract (16.2% of all the contracts)[1].” – not clear

Response 3: this sentence has been amended to

The current recovery and growth of the Spanish economy encouraged immigrant arrivals to recover to pre-crisis numbers, and immigrants now make up 16.2% of the workforce [2].

Point 4: The authors should explain why they focused on this particular cohort (Colombian and Ecuadorian). Also, earlier in the text they provided 3 alternative hypotheses for the decline of migrant health. Do the authors plan to test any of them in this study?

 Response 4: This is the only cohort study of immigrants in Spain that permitted this analysis.  We were not able to use the Spanish Working Conditions survey because it has a cross-sectional design and too few immigrants in the study.

The theories we presented were explanations and not hypotheses. We examine the first one (acculturation).

Point 5: Methods:  Operationalization of “good/poor mental health”. Why was the cut-off score set at 3? Was it based on the prior literature?

 Response 5: We apologize for having missed the bibliographic reference (Jackson, 2007) that established the cut-off score as three. We have added it in the text as well as in the reference list.

Point 6: Only 130 participants (of the 473 interviewed at baseline) were employed and in good mental health at baseline. That is only 27% of the sample. Was this mostly driven by the employment status or by the poor health of the respondents? Could the authors comment on the high prevalence of poor health in their sample and compare it with what was reported in the literature?

Response 6: We have added in the paragraph (2.3 Analysis) that there were 324 (111 Spanish and 213 immigrant) workers in employment at the baseline interview. So, the prevalence of good mental health was 42% (i.e. 186/324), over all workers. This figure is similar to that found in other studies that have been mentioned in the discussion – see references below.

1.      .Robert, G.; Martinez, J. M.; Garcia, A. M.; Benavides, F. G.; Ronda, E., From the boom to the crisis: changes in employment conditions of immigrants in Spain and their effects on mental health. Eur. J. Public Health 2014, 24, 404-409.

2.      Antecol, H.; Bedard, K., Unhealthy assimilation: why do immigrants converge to American health status levels? Demography 2006, 43, 337-360.

3.      Subedi, R. P.; Rosenberg, M. W., Determinants of the variations in self-reported health status among recent and more established immigrants in Canada. Soc. Sci. Med. 2014, 115, 103-110.

4.      La Parra-Casado, D.; Stornes, P.; Solheim, E. F., Self-rated health and wellbeing among the working-age immigrant population in Western Europe: findings from the European social survey (2014) special module on the social determinants of health. European Journal of  Public Health 2017, 27, 40-46.

5.      Gkiouleka, A.; Avrami, L.; Kostaki, A.; Huijts, T.; Eikemo, T. A.; Stathopoulou, T., Depressive symptoms among migrants and non-migrants in Europe: documenting and explaining inequalities in times of socio-economic instability. Eur. J. Public Health 2018, 28, 54-60.

Point 7: L106: typo “34”

 Response 7: Done: 34-41

Point 8: Why did the authors decide on the following categories: 1-10, 11-15, and above 15? How many respondents were in each of the cat:. 


Response 8: We have added a paragraph clarifying the point mentioned by the reviewer:

To establish categories 1-10, 11-15 and> 15, the following strategy was used: First, the sample was split into quartiles, obtaining groups 1-10 (n = 19), 11-13 (n = 22), 14-15 (n = 48) and> 15 (n = 13). Secondly, the incidence of common mental disorders was obtained in each of the groups. The incidence was 15.8%, 22.7%, 25.0% and 46.2%, respectively. Due to the similarity in the incidence of groups 11-13 and 14-15 and the small sample size, we decided to create a single group of 11-15 years. Therefore, the final number of categories were 1-10 (n = 19), 11-15 (n = 70) and> 15 years (n = 19).

Author Response File: Author Response.docx

Reviewer 3 Report

The theme is of great relevance for public health policy decision-making. It is a very positive point of the article, which increases its relevancy for publication. However , some changes must be made , as specified in the file attached.

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 3 Comments

Point 1: ABSTRACT: Objectively reproduces the content of the work, although it does not clearly state the methodological assumptions.

Response 1: We are not clear what you mean by methodological assumptions

Point 2: INTRODUCTION: in L-35 there is a comparison of the percentage of migrant individuals, in relation to the total population. However, the two years cited are 1998, which is why the setback loses its meaning.

Response 2: Fixed - thanks

Point 3: In addition, the transition between the problem contextualization and the deepening of its unfoldings happens in a somewhat hurried way, which, in addition to the small grammatical errors contained in the text, especially in relation to the use of connectives, makes the reading less fluid.

Response 3: We are not clear what you mean by deepening of its unfoldings. Author 3 is a native English speaker.

Point 4:  METHODS: In 81-83, the age criteria established for family members eventually creates confusion. It happens because, if the children have to be between 12 and 17 years old, the minimum age of the responsible adult cannot be 18 years old, because, if so, he/she could not be the parent. In this way, it is necessary to clarify if this adult has to be specifically the parent.

Response 4: the following sentence has been included.  So for this study we included only the 473 adults aged ≥18 years.

.

Point 4: RESULTS: Although the result is mostly adequate, it is necessary to clarify in wich way the found ORa values support the present study hypothesis, since the values are ORa <1, which points to the existence of a protective factor - issue not addressed by the authors.

Response 4: We assume that the readers of this journal understand how to interpret an adjusted odds ratio.

If the editor insists that this is included in the paper – we suggest using “An ORa greater than one indicates higher risk whilst and ORa less than one equates to reduced risk”.

Point 5: DISCUSSION: In general, the specificity of the immigrant’s origin country is not sufficiently addressed, since the only aspect analyzed is the language factor, which can be applied to several Spanish - speaking countries.

Response 5: in the discussion – lines 191 to 205 we discuss several explanations (language, nationality, socioeconomic status, migration status) as explanations for a lower risk of common mental disorders in immigrants.

Point 6: It would be important to have presented more characteristics that justified the choice of Ecuador and Colombia.

Response 6: In the methods section we have included the following sentence.

Migrants from Columbia and Ecuador are the most highly represented nationalities in Spain, from Latin America.

Point 7: In addition, there was no mention of the occurrence of the protective factor for all the variables involved in the study.

Response 7: Our objective was to analyse the incidence of common mental disorders by  migration status and taking into account the time of residence. The other variables involved in the study were considered confounding covariates. We applied an approach¹  to control for all the other explanatory variables and just focus in our objective.

1.     Twisk JWR. Applied Multilevel Analysis: A Practical Guide for Medical Researchers Cambridge University Press. 2006.

Author Response File: Author Response.docx

Reviewer 4 Report

This is an interesting paper on a very important topic in migration studies: the mental health of migrants, in comparison with natives. However, some changes are needed in order to improve the paper.

1) There is a lot of international literature on Latin Americans as selected immigrants in the country of destination. Part of this selection is not attributable only to the healthy migrant effect or to the salmon bias, but also to cultural aspects. I think that previous studies on this topic should be mentioned.

2) I suggest to show the odds ratios of all the independent or control variables considered in the multivariate analysis and not only those regarding the duration of residence. 

3) Among the limitations of the study, you should mention also the sample size, which is quite limited. 

Author Response

Response to Reviewer 4 Comments

Point 1:  There is a lot of international literature on Latin Americans as selected immigrants in the country of destination. Part of this selection is not attributable only to the healthy migrant effect or to the salmon bias, but also to cultural aspects. I think that previous studies on this topic should be mentioned.

Response 1: We are not sure what you mean here? Do you mean that Latin American immigrants are selected into the host country because of cultural aspects or do you mean that our findings (of reduced risk of poor mental health) can be explained by cultural aspects?  There are papers on Latin American migrants in the literature but most of these relate to Latin Americans in the USA, where the situation is different than in Spain. For example, communicaton for LA migrants in the US is more difficult than in Spain. It is easier for migrants to become a Spanish citizen than is the case in the US. In Spain the main religion is Catholic, as is most common in LA. In the main, migrants from Ecuador and Columbia have a similar culture and diet to the Spanish and Spain has a national health service.

Point 2: I suggest to show the odds ratios of all the independent or control variables considered in the multivariate analysis and not only those regarding the duration of residence.

Response 2: Our objective was to analyze the differences in the incidence of common mental disorders by migration status and time of residence. The other explanatory variables are considered as confounding covariates. Showing a table with all the associations of covariates (that is, including confounding factors) dilutes the main result and moves away from the objective of the study. The two references below describe the method we applied.

1.         Twisk JWR. Applied Multilevel Analysis: A Practical Guide for Medical Researchers Cambridge University Press. 2006.

2.         Benavides FG, Benach J, Muntaner C, Delclos GL, Catot N, Amable M. Associations between temporary employment and occupational injury: what are the mechanisms? Occup Environ Med. 2006 Jun;63(6):416-21.

Point 3: Among the limitations of the study, you should mention also the sample size, which is quite limited. 

Response 3: The following sentence on sample size calculation has been added to the methods section.

Sample size was estimated with a Poisson approximation. Three hundred and fifty nine immigrants and 107 natives were required to detect a minimum relative risk of 1.5 (based on earlier work), assuming an alpha of 0.05 and beta of 0.2 and an attrition rate of 15%. Assuming four family members per family (an estimated two children per family, based on the number of children per woman of foreign origin in Spain 1,54 [2]) and rounding upwards gives 180 immigrant and 50 Spanish families in each city.

An also in the conclusions:

Bearing in mind that these findings come from a study with a small number of participants policies aimed at preserving the mental health of new arrivals and improving the mental health of Spanish-born and longer duration of residence immigrants are warranted

Author Response File: Author Response.docx

Round 2

Reviewer 4 Report

The paper can be published in the present form.