Mixed Method Synthesis of Cognitive interviews with Kenyan peripartum adolescents: Testing EPDS and PHQ questionnaires in Nairobi primary care settings

Background: Cognitive interviewing is a well-recognized methodology to identify clinical and commonsensical relevance of mental health questionnaire items by our research participants. Depression is amongst the most common condition impacting pregnant and parenting adolescents in sub-Saharan Africa (SSA). In Kenya, studies have reported depression prevalence estimates of 12-50% in peripartum adolescents. While young people prefer using English, there has not been enough data to point to how well they respond to Kiswahili translations of the commonly used tools. Method: Thirty-two participants between ages 14-18 years were approached and through informed consent for them to participate. We used Edinburgh Postnatal Depression Scale (EPDS) and Patient Health Questionnaire-9 (PHQ-9) in English and Kiswahili versions to carry out the interviews and were coded along four domains of comprehension, retrieval, judgement and patient response. Scores were tabulated for all participants and simple reliability analyses were offered. The interviews were discussed within the team and tables were shared between three raters for further consensus. The interview transcripts were analyzed manually using thematic analysis. Discussion: We found that adolescents had challenges in fully understanding items in both set of tools however, EPDS was better received than PHQ-9. Psychometrically, EPDS English version and Kiswahili versions fared better than the PHQ-9 English and Kiswahili versions. PHQ-9 presented considerable issues with regards to semantic clarity however had simpler response options in comparison to EPDS which was more experientially appropriate, but response options were not simple. On our thematic analysis we felt that the adolescents were signicantly challenged by the new, unanticipated pregnancy or motherhood experience. Poverty, poor partner support, discord with parents and distress in the family, and traumatic end of supportive and helpful relationships were the notable problems. Conclusion: While sensitive cross-cultural translations for the commonly used open access depression tools is critical, it is also pertinent to understand whether these are developmentally appropriate. Our participants had multiple psychosocial and material challenges that necessitate measures sensitive to their age, social context and health experiences. Additionally, we felt that both the tools may not be suitable for self-administration and may need involvement of a case worker or community health worker.

females, the burden of depression remains 50% higher for females than males (10). Currently, depression is the second leading cause of Disability-Adjusted Life Years (DALYS) following anxiety as the leading cause (1).

Studies on Local Burden Associated with Adolescent peripartum Depression
In a study conducted in Nairobi by Osok et al to screen for depression among women attending antenatal clinics for a sample size of (n=176); it was found out that 58% showed likelihood of depression using EPDS. This was then con rmed using PHQ-9 and 33% (n=58) positively screened for clinical depression (11). In Kenya a study conducted by Ongeri et al found that the prevalence of depression using EPDS was 18.7% (95% CI: 13.3-25.5) (12). Another study conducted in Addis Ababa, Ethiopia by Biratu et al on antenatal depression in pregnant women using EPDS scale (n=393), found prevalence of antenatal depression to be 24.94%(95% CI: 20.85-29.30%) (13). In another Ethiopian study conducted by Dibaba et al pregnant women (n=627); the prevalence of depressive symptoms during pregnancy on EPDS scale was 19.9% (95% CI: 16.8-23.1%) (14) and Govender et al in South Africa found (n=326) the prevalence of depression among the pregnant participants was found to be 15.9% using EPDS (15).

Cognitive interviewing as a method and theoretical approach
Cognitive interviewing of patients for whom the questionnaire is intended proved a valuable tool to increase its face and content validity (16). It is underpinned by the assumption that respondents are able to provide verbal reports of their thought processes (17). Cognitive Interviewing (CI) refers to a method used in research to help the researcher identify various sources of confusion in assessment items and to assess validity evidence on the basis of content and response processes (18). The process involves scrutinizing the question-response to determine how well a question performs in capturing a particular concept when asked to respondents. The goal is for the EPDS 10 items and PHQ-9 items to capture the scienti c intent of the question and, at the same time make sense to respondents.

Method Participants, Setting and Ethical Clearance
The study was approved by Kenyatta National Hospital/University of Nairobi ethical review committee (approval no. P694/09/2018). Approval was also received from Nairobi County Health Directorate (approval no. CMO/NRB/OPR/VOL1/2019/04) and subsequently permit from Kenyan National Commission for Science, Technology and Innovation (NACOSTI/P/19/77705/28063) was sought.

Tools, their translations and psychometric properties
This study used a mixed method design to interpret data collected in two health care facilities, in Nairobi County-Kenya in 2019. Data collection was through use of questionnaires interpreted quantitatively and semi-structured interviews (qualitative) capturing their responses, comprehension, retrieval and judgment of the questions in the quantitative questionnaires pinned in cognitive interviewing theory. The tools used were Edinburg Postpartum Depression Scale (EPDS) and Patient Health Questionnaire (PHQ-9) each with both English and Kiswahili Versions on 32 participants from July-November (5 months).
In the United States it has been applied on various populations such as the women and Spanish women. In this study, a sample of 1516 postpartum women completed the EPDS and the Beck Depression Inventory (BDI). Severity ranges in this study matched those of the Beck Depression Inventory with majority of the women between a score range of 7 and 13 (minimally depressed) likely to be mild depression with other tools measuring depression severity. This classi cation allows proper referral and treatment decisions to be made (20).
The EPDS has been validated in different cultures showing its importance in mitigating long lasting effects for mothers, their children and society at large (21). It has been widely tested globally including countries in Sub-Saharan Africa (SSA) (22,23). Translations in Kiswahili developed in Kenya, where EPDS was translated into Kiswahili and it is the version that is being used here (24).
In South Africa it was found to be reliable for criteria-based clinical depression disorders diagnosis, it is also good for severity check and perfect for clinical and research use. The study conducted a validation study of an interviewer-administered brief screening tool for depression, in a high HIV burden, low literacy PHC population in Johannesburg, South Africa (26). The results provide insight on how we can screen for and prioritize often overlooked and highly prevalent depressive symptoms in low-resource PHC settings in sub-Saharan Africa. This study provides a framework for implementing depression screening programs in resource-scarce sub-Saharan African contexts and establishes the PHQ-9 as a useful screening instrument in these settings (27).

Data collection Procedure
It encourages an in-depth exploration of particular concepts, processes and patterns of interpretation. To achieve this, the process of cognitive testing analyses the question-response process in four stages: i. Comprehension: This is a stage where the respondent interprets the question. Di culty at this stage may re ect unfamiliar or complex expressions or words or unclear concepts.
ii. Retrieval: At this point respondent searches memory for answer. Recall di culty can be a challenge at this stage.
iii. Judgment: The respondent estimate response. Di culty in estimating response may re ect bias or sensitive question.
iv. Response: The respondent provides answer in the format requested. The answer provided may expose issues of incomplete response options.
The aim is to understand how each participant interprets the assessment items and selects responses. Due to the personal nature of cognitive interview, in that it involves asking the respondent what they are thinking or feeling when answering a question, it will therefore, be made clear to respondents when they have volunteered for a cognitive interview what it involves in order to maximise cooperation and yield from the process. The research team will also attend to any issues with sensitivity of the questions or poor comprehension of the domains to be assessed.
Cognitive interviewing as a method explores how the researchers' and participants' understanding of the constructs intersect. It allows an opportunity to scan the participants' perceptions, fund of knowledge, comprehension and at the same time gives unique information on the acceptability of the ideas tested in a tool. It allows cross-cultural, developmental (as we deal with adolescents) validation of depression constructs.
The participants were taken through a brie ng on the objective of the interview and their written and oral consent taken to record their experiences. Participants were assured that they could withdraw when they felt like discontinuing and were assured of assistance if screened with high depressive symptoms. Referrals were made to the psychiatric nursing o cer at the health facility and individual follow up done by the clinical psychologist research assistant.
The protocol was developed by the rst author and the two research assistants MK and VN interviewed and collated data on the tools. A structured worksheet was available for research assistants to transcribe the participants' response verbatim. These notes were translated by the interviewing research assistants. The study team reviewed cognitive interview transcripts, discussed and addressed items requiring revision.

Data analysis
At rst all the scores were compiled for a brief summary score report and reliability analysis for the two scales and two translated versions carried out. The cognitive interviewing was sequenced in a four domains assessment which was collated by the study team. Each participant was offered both tools but they could choose, English or Kiswahili version and for each item they were asked what the item a) meant and how well they understood the question b) what did they think it asked about and c) what their own response was; including being administered the questionnaire to map their depression experiences and severity. This data was descriptively presented to re ect participants' understanding of the tool items. Finally, a thematic analysis informed qualitative synthesis of the core ndings of these interviews was carried out to present gaps in our understanding of depression in young adolescent girls, their own life experiences and clinical signs and symptoms that depression and other mental health interventions need to target.

Results
The key ndings are broken down into three sections. First, is the presentation of cognitive interviews along the four domains of comprehension, judgement, retrieval and experiential response to items described earlier and along with it we present their summary scores and a descriptive report on item level reliability analysis computed through Cronbach alpha. Finally, we offer some qualitative analyses of what depression meant for our sample and the kinds of associations to depression that young girls brought out in relation to the pregnancy or motherhood experience. We interviewed thirty-two female participants between ages 14-18 years. Half of the participants were pregnant, and the other half were postpartum and twenty-one of these participants were living on their own or with parent (in the absence of a partner) and the remaining were married. See table 1 for sociodemographic characteristics of the participants and  table 2 for depression cut off scores for both tools.

Cognitive interviews summary
We had eight participants interviewed using EPDS Kiswahili and another nine using EPDS English version. Nine participants were interviewed using PHQ-9 English and six using PHQ-9 Kiswahili version. As described in the methods section, the participants were scored for their ability in comprehension, retrieval, judgement and response quality. We were also scoring these participants on the questionnaire using the formal scoring. See table 3 that looks at cognitive interview rating for PHQ-9 and EPDS English and Kiswahili versions.
On PHQ-9 English interviews (n=9) we found that our participants had a di cult time following the item 1, 'little interest or pleasure in doing things', with explanations such as, 'you could want to do something but you don't have the support needed, When one lives in a mistreated environment'. On item 2, 'feeling down depressed and hopeless', one of our participants doesn't understand the item, she says 'doesn't comprehend it all "feeling down" is hard'. On item 4 'feeling tired or having little energy', most of the participants understood the item however there were odd one or two responses where there was some unclarity around comprehension. Responses such as 'she feels lazy/feeling "maybe I am sick" she forces self to go to work' leave us wondering whether this 15 year old with 7-month pregnancy is able to grasp the item fully. Item 5 was well understood by all. On item 6 'feeling bad about yourself-or that you are a failure or have let yourself or your family down' some responses like 'blames oneself, feels she is a let-down' does convey that there is an overwhelming sense of resonance with the theme that the item conveys. On item 7 around 'trouble concentrating on things, such as reading the newspaper or watching television', there were at least two participants who had issues with comprehending the question with responses such as 'not being able to listen to news' or another 15 years old saying that she didn't get the question. On item 8, 'moving or speaking so slowly that other people could have noticed. Or the opposite-being so dgety or restless that you have been moving around a lot more than usual' one of our 18 years old mother of one-month old child responded 'when unhappy her voice goes down', another 6 months pregnant 16 years old said that 'If you change hormones to walk or talk with say bitterness, anger'. On item 9 about 'thoughts that you would be better off dead, or of hurting yourself', almost everyone had a good sense of what it meant.
On PHQ-9 Kiswahili cognitive interviews (n=6), we found that two of our participants had a hard time following the question. With one 18 years old with a 10month old baby saying, 'I have not understood the question' and another one having di culties understanding the response items. On item 2, a 17 years old pregnant participant does not follow the item despite repeating and another participant 18 years old mother responds, 'if you are going through a lot of trouble'. Item 3 was well-understood by all our participants. On item 4, one of our 18 years old young mother's feedback was that 'you feel tired after doing a lot of work' and such responses can be perplexing as one may be offering one's lived experience than the understanding of the question per sae. Another 17 years old 6-months pregnant participant felt that she didn't understand the question and needed it to be explained. On item 5 one of our 17 years old participants shared her understanding that 'maybe if you feel like eating' which again did not capture the actual meaning of the item. One of our 17 years old 7-months pregnant participant had problems following the item 6. Another 17 years old 6 months and 18 years old young mother responded to the item as 'maybe your parent asks you for something and you don't have; you feel bad', and 'for example parents ask for money at the end of the month but you don't have,' another 14 years old 10-months pregnant adolescent said 'means that you feel bad and not happy about your family and you think of bad things'. On item 7, one of our 16 years old 7-months pregnant adolescent responded that 'paying attention while doing something or reading' and the 18 years old adolescent mother said, 'it's like reading a magazine reduces stress'. A 17 years old participant wanted this item to be explained again. Another 18 years old with one-year old baby made an interesting comment 'I don't have a TV and also don't read newspapers.' On item 9, one of our 16 years old 7-months pregnant participant's interpretation was 'people realizing that you have changed a lot. Looking as if you are absent-minded'. One of our 18 years old young mother didn't understand the item while another 18 years old young mother said that 'I walk slowly because walking fast can make you get hit by a motorcycle'. On item 9, one 17 years old 6-months pregnant adolescent said that 'I have understood but I don't know how to explain' and other than that all participants understood the meaning of the item well.
On EPDS English version (n=9), on item 1, 'I have been able to laugh and see the funny side of things', one of our 18 years old with 2 months pregnancy, responded 'I have been able to laugh and see the funny side of things' though another 16 years old 5-months pregnant adolescent did not follow the item despite being reminded a number of times. One 14 years old 6-months pregnant adolescent responded '"seeing the funny side of things" is di cult to understand' and another 14 years old responded 'I can laugh but sometimes I am angry'. On item 2 'I have looked forward with enjoyment to things.' Our 14 years and 6-months pregnant participant says, 'if I can see and enjoy' and the 16 years old 5-months pregnant participant does not understand the question. Another 14 years old 6-months pregnant says 'waiting for the pregnancy with joy' and another 15 years old with 10-months old baby says, 'due to baby responsibility she has to work hard and have hope'. On item 3, 'I have blamed myself unnecessarily when things went wrong', one of our participants says, 'I hated myself', another 16 years old 5-months pregnant participant does not understand the question even after multiple repetitions. On item 4, 'I have been anxious or worried for no reason', one of our 18 years old participant who is mother of a two-year old says, 'when thinking of something that will hurt you but others see success,' and our 16 years old 5-months pregnant and another 14 years old 6-months pregnant participants 'did not know the meaning of "anxious"' . On item 5, 'I have felt scared or panicky for no very good reason,' our participants interpreted it as, 'you might want to do something but fear of how people will react; yet others see that you are able to succeed in doing it'.
On item 6, 'things have been getting on top of me', our 14 years 6-months pregnant adolescent interpreted as, 'things can come at my time', while 16 years old 5-months pregnant adolescent thought it was a di cult question to follow, another 14 years old 6-months pregnant participant found it di cult as she explained it as, 'things that are continuing to happen to me'. Another 15 years 10-months pregnant participant explained it as, 'if she had a partner, she would be able to carry the big load rather than alone'. On item 7 'I have been so unhappy that I have had di culty sleeping', one of our 17 years old 2-months pregnant participant interpreted this as, 'if she has a problem it means that she has to step up' and two other participants 16 years old 5-months pregnant and 14 years old 6-months pregnant did not understand the question despite being repeated. On item 8, 'I have felt sad or miserable', two of our 18 years old 2months pregnant and 14 years old 6-months pregnant participants were not able to understand the word "miserable." Another 17 years old mother of 2months old baby associated it with 'being helpless' and two participants, one 14 years old 6-months pregnant and another 16 years old 5-months pregnant interpreting it as, 'I can lack happiness'. On item 9, 'I have been so unhappy that I have been crying', one 16 years young mother said 'she reports crying at rst but now she has accepted', and another 14 years old 6-months pregnant participant interpreted as, 'if I can be unhappy and cry,' and another 16 years old 5months pregnant participant said she understood but didn't respond. On item 10, 'the thought of harming myself has occurred to me', most participants understood it however our 14 years old 6-months pregnant participant said 'that one is di cult', another 16 years old 5-months pregnant participant understood but did not respond.
On EPDS Kiswahili version (n=8), on item 1, one of our 16 years old 8-months pregnant participant said, 'if you have laughed because you have seen things', one 17 years 2-months pregnant participant did not understand. Another 15 years young mother of one also did not understand the item. On item 2, one of our 17years old mother of 8-months old child also did not follow the question and another 17 years old 2-months pregnant participant also did not follow the question. Another 15 years old young mother said doesn't understand the meaning of "I have looked forward" and the 18 years old 8-months pregnant said, 'it means if I have looked forward to things with enjoyment'. On item 3, one 16 years old 8-months pregnant interpreted it as, 'if you were not used to blaming yourself', another 17 years old 3-months pregnant participant did not follow the question and a 15 years old young mother said, 'if you insult your parents'. On item 4, our 16 years old 8-months pregnant participant said, 'if you never liked being worried', another 17 years old 3-months pregnant participant did not understand. On item 5, our 16 years 8-months pregnant participant said, 'In the past you were not scared or anxious', however others did understand the item well. On item 6, one of our 16 years old 8-months pregnant participant said, 'in the past you never used to face issues but now you are facing them', and another 17 years old mother of 8-months old said, 'it means failure to understand well'. Another 17 years old 3-months pregnant participant did not understand the item. On item 7, one participant a young 15 years old mother of one-year old child said, 'if your loved-one dies'. On item 10 our 15 years old mother of one did not follow the question. Some of our participants had a hard time following the questions especially on the Kiswahili versions of the tools (which they preferred over English). Often times they would say that the question is hard to follow. For example, when asked if the participant experienced same levels of enjoyment in things she used to like doing before, one of our 17 years old 2-months pregnant participant said 'that question is hard', she repeated this for couple of other questions. Another 17 years old 3-months pregnant participant when asked if she blamed myself unnecessarily when things went wrong from EPDS said that she didn't understand the question and that it was hard.
Psychometric analysis of the two tools with two translated versions Tables 4 and 5 offer reliability analysis of the two tools. We can see that for PHQ -9 combined the overall Cronbach's α for internal reliability is 0.56 and for EPDS is 0.76. On EPDS, 5 participants had a score of over 14 indicating moderate depression, on PHQ-9, 6 participants who scored under moderate depression category and 5 participants who scored 15-19 moderate severe depression. Our sample size is certainly limited and therefore it is imperative to interpret these ndings cautiously. Mean score of EPDS English version is 12.89 (SD=5.71) and Kiswahili version is 9 (SD=6.55) and mean score on PHQ-9 English version is 12.44 (SD=4.69) and Kiswahili PHQ-9 version is 11.33 (SD=4.08).

Thematic analysis of the participant interviews
The interviews were at another level very poignant and painful reminder of how deeply affected these adolescents were in some ways by the pregnancy and unintended/ unplanned motherhood.

Manifestations of depression
Our participants shared through these interviews several feelings and experiences associated with being, down dejected and hopeless. Almost always there were adverse family experiences and lack of provision for their basic needs that contributed to feeling dejected.
Yeah you cry and ask yourself why did it have to happen to-why is it only me that was supposed to be in this situation; so I ask myself several questions I even lose several things in life, education. Thoughts of harming myself have come to me several times. (18 years old with 2 months old baby).
You can just be in deep thoughts thinking a lot of things, maybe you are stressed and you are unable to fall asleep…. (18 years old mother with 8 months old baby, PHQ-score of 12) The second participant often contemplates suicide when her father is rough with her and the mother is pressing her to stay with the baby rather than take up a job and fend for herself and family.
When you wake up feeling that way and the parents are around, they start quarrelling saying that you are a girl and you don't want to work you have left all the house chores for other siblings. At times feel very tired and hungry -sometimes there is food sometimes there isn't. I normally feel that I have let them down because they expected me to at least to raise their living standard in future but instead I got pregnant. (16 years old with a 4 months old baby and a PHQ-9 score of 18) Depression is lacking peace… you feel bad about your body or it's like you have let yourself or your family down. … I was in school when I got this child so my mother just wasted all her money taking me to school. I just feel bad because right now I am at home and all my sisters are at school continuing with their studies so I tell myself if I wouldn't be in this situation I would be in school. (18 years old mother of 2 months old infant, PHQ-9 score of 13).
Similar sentiments around being disenfranchised were shared by another participant.
That feeling you get when wronged, you don't want to talk to anybody and just want to be left alone I do feel bad. At times I even contemplate suicide. (16 years old mother of 4 months old infant, scored 18 on PHQ-9) Absence of social support and loneliness was also alluded to by a participant.
Because the child's dad does not support me and I am suffering alone. I have contemplated suicide. I usually sit down and correct myself; I talk to myself and say that this is not the best thing to do. There are many challenges…..lacking a person to talk to sometimes (17 years old mother with EPDS score of 19).
These experiences of not nding support are also complex and also worthy of further exploration as they hinge on several challenging experiences of being left alone, not having anyone to talk to, going hungry, lacking provision of basic needs etc. The descriptions make it seem that the experience of unintended and unplanned pregnancy is a signi cant jolt for more or less every adolescent.
Acceptance of pregnancy While accepting the pregnancy and receiving support from the parents to help deliver the baby, the adolescent feels that there will be signi cant trade-offs between returning to school, nding job and caring for the baby. Even though the pregnancy is accepted helplessly as a fact of life, the implications of longterm responsibility and support that an adolescent has to vouch for, is something that young girls seem to struggle with tremendously. Support and stresses around adolescent's caregivers especially mother It means the way I am and the way I was, the level of happiness before and now, if I am just happy just the way I used to be, I was very happy when I was pregnant and up to now it is not gone especially when I see my baby, it makes me feel happy because I know even though I am raising the baby alone I know one day she/he will help me (17 years old with 10 months old infant, score of 13 EPDS) 'Yes. Because sometimes I just ask myself 'now where is my life heading to?', where I am today and where I will be, I sometimes try to give myself hope and then I feel that maybe I am giving myself hope yet I will not reach where I want to be…. It is like when you over-think and feel like life has become di cult and you cannot take it anymore so you feel like the only solution is to take your life to avoid troubles, but again when you think about your baby you feel that 'who will I leave him with?' and you also don't want to leave that burden to your mother, because you will also leave her with stress and she can even die.(17 years old 10 months old) problem of I mean many troubles whenever you are stressed it leads somebody to lose whatever -there is something you just lose; it is like you become insane. You lose a lot -cannot go to school… because you might nd that even in school people are laughing at you, you regret so it makes you feel like you are stigmatized. My self-blame comes from due to peer pressure about others and my friends they led me into a wrong way then I fall out [Inaudible] I used to blame myself….now I have returned to school (18 years old mother with 2 years child, 16 score on EPDS) it reached a point when my mother, you know my mother left us when we were still very young. When I was three years old when she died, so we were left with my grandmother, then my father was arrested, so we went through a lot, my grandmother's leg broke, so our grandfather took the initiative of feeding us, clothing us. Taking us to school; either we go or we don't because we have been sent away due to lack of fees and we are just at home. It reaches a point I could start to cry and say [Inaudible] 'why has mum left us?' Our father lived in prison about four years (18 years old mother of 8 months baby, PHQ 9 score of 12).
Demanding multitasking and early adulti cation Stress at home yeah; you may have gone to school then she tells you, you know she remains back; she doesn't go to work then you tell her 'mum since you are not going to work can you stay with the child for me for a while,' then she will refuse, even if it is Sunday. I may come back from school in the evening then you nd that everything is waiting for you there so (18 years old with 2 years old baby, 16 on EPDS) Maybe you had something urgent that you wanted to do but you have not done it because maybe you have a lot of work to do, so you will be forced to be more faster to be able to handle other tasks, or maybe you have a lot of stress (16 years old with 4 months baby, 18 on PHQ)

Discussion
The paper has adopted a two-pronged approach of introducing ndings of cognitive interviews from the two sets of tools, brie y reviewing their psychometric properties using Kiswahili and English language versions and developing a thematic analysis on the experiences around various questionnaire items. We were hoping to summarize the life experiences of our participants as well as experience and process of conducting a cognitive interview exercise through this approach.
Psychometrically, we think that EPDS is a better tool than PHQ-9 for our adolescent sample for two reasons-one, the latter is more congruent with the peripartum context and second, the questions are organized more sensitively. The Kiswahili translation was signi cantly better than the PHQ-9 translation. On both tools we can see several items where the item reliability scores were poor and can see that PHQ-9 had many more items with poorer alphas. We think that the factor structure of the tools may be different for pregnant and parenting adolescents than the adult samples. This warrants further analysis that is currently outside the scope of this paper and we acknowledge that our sample is limited to make clear conclusions.
In Kenya Cognitive testing of the PHQ-9 has been conducted for depression screening among pregnant and postpartum women (28). Results showed that most participants preferred to answer the PHQ-9 in Kiswahili (N = 15; 52%). Among the 20 interview participants, 12 (60%) had scores ≥5, indicating depressive symptoms. Overall, participants found the scale acceptable as an interviewer-administered tool. Participants reported few problems related to comprehension but had di culty answering items not relevant to their lives (e.g., "watching television") and double-barreled items (e.g., "poor appetite or overeating"). They were hesitant to endorse items related to "duties as a wife and mother" and suicidal ideation. Most participants had di culty distinguishing between response options of "several days" and "more than half the days".
Cognitive interview ndings do make us feel that the distillation of actual experiences and meaning of those states of mind and bodily reactions may not completely t with each other. However, at the same time separating the two would be di cult. Both tools were better received for most parts when delivered in English as young people felt more comfortable in English. Very chaste Kiswahili may no more be used in urban informal settlements where 'sheng' a hybrid of Kiswahili, English and colloquial dialects are used.
It is sometimes hard to be an objective witness of one's distress while undergoing in that moment. This is something that trauma researchers know well in their work around testimonies of trauma. At times these cognitive interviews were akin to testimonies of pain and sadness of life circumstances and adverse choices one has made. In almost all statements that the participants offered their understanding and experience, we noted the overwhelming distress, helplessness, sadness and sense of isolation they went through. There was early adulti cation and end of school education, very painful reality of looking after a new being and not having time or resources to become a full time carer. The interview transcripts were a testament to how stoic adolescents were around particular questions of whether they have let others down, feeling hopeless and that their life is not worth living. Their pauses, delayed responses and at times asking for the question to the repeated again felt as though these questionnaire items were like frozen metaphors for their state of mind and existence.
We found interesting themes like participants wanting more attuned care and support from their mothers and at the same time knew how hard things were going to become for the adolescent's mother. This duality was very striking. The interpersonal web of challenges around adolescent pregnancy and motherhood is quite signi cant. The missing supportive male caregiver and absence of a wiling and responsible partner/boyfriend cannot be underscored. The participants also appeared to become more aware of their state of mind as the questionnaire items were presented. In a context where mental health is poorly understood, these questionnaires become an odd encounter of coming face to face with all cognitive processes and emotions that must underlie adolescent experience of encountering unplanned motherhood.

Limitations
We did not administer the four sets of tools on all participants. It would have been di cult from a design as well as measurement point of view repeating similar kinds of questions on participants raising even ethical issues of exposing them to further duress.

Conclusion
For pregnant and parenting adolescents in low resource settings or in LMICs it would be important to consider using the self-administered depression or mental health questionnaires carefully. It may be important to have group assessments facilitated by an experienced social worker, health worker with training in mental health or by a psychologist. We need to make response items of these public mental health tools more responsive to adolescent and less literate populations. Kenyan pregnant adolescents and adolescent mothers demonstrate signi cant mental distress and depressive symptoms however the precise assessment of the extent of their depressive symptoms cannot be reliably ascertained given the issues that have been presented in this paper.

Declarations
Ethics approval and consent to participate: the study which has been approved by Kenyatta National Hospital/University of Nairobi ethical review committee (approval no. P694/09/2018). The study received approval from Nairobi county health directorate (approval no. CMO/NRB/OPR/VOL1/2019/04) and approval from National Commission for Science, Technology and Innovation (NACOSTI/P/19/77705/28063). The written informed consent to participate was sought from all study participants.
Consent for publication: all study participants would be asked for their written consent to publish the ndings of this work. The key study participants and their family member data would be anonymized, and the personal information would be kept con dential Availability of data and material: NIMH funded research requires the data to be made available in publicly available. The data from this study would be made available on repositories such as Clinicaltrials.gov, RDoCdb, FIC webpage. All the personal information would be de-identi ed, and the data put on excel sheets for research use. Some interview transcripts would be kept by the researcher and may be shared on reasonable request.      1  0  0  2  1  1  2  2  2  2  2  2  2  2  2  2  2  2  2  2