An Adapted Self-Screening Tool for Peripartum Cardiomyopathy in Haiti

Abstract

Background:

Peripartum Cardiomyopathy (PPCM) is considered rare in the United States, however the literature notes that the disease has a higher prevalence in developing countries such as Haiti. Dr. James D. Fett, a U.S. cardiologist, developed and validated a self-assessment measure for PPCM in the United States to aid women to easily differentiate the signs and symptoms of heart failure from those related to a normal pregnancy. Although this instrument was validated, it lacks the adaptation necessary to account for the language, culture, and education of the Haitian population.

Objective:

The aim of this study was to translate and culturally adapt the Fett PPCM self-assessment measure for use among a Haitian Creole speaking population.

Methods:

A preliminary Haitian Creole direct translation was developed from the original English Fett self-test. A total of four focus groups with medical professionals and 16 cognitive interviews with members of a community advisory board (CAB) were conducted to refine the preliminary Haitian Creole translation and adaptation.

Results:

The adaptation focused on incorporating cues that would be tangible and connected to the reality of the Haitian population while maintaining the intended meaning of the original Fett measure.

Conclusion:

The final adaptation provides an instrument suitable for administration by auxiliary health providers and community health workers to help patients distinguish symptoms of heart failure from symptoms related to normal pregnancy and further quantify the severity of signs and symptoms that might be indicative of heart failure.

Introduction

Peripartum Cardiomyopathy (PPCM), a form of heart failure that manifests among pregnant and early postpartum women, has a prevalence of one case per 300 to 400 live births in Haiti (1). With this disease, the heart chambers become enlarged, thereby weakening the muscle and making it difficult to effectively pump blood (2, 3). The common signs and symptoms of PPCM are often similar to those attributed to a normal pregnancy, therefore the condition is often unnoticed or misdiagnosed. Though the pathophysiology of PPCM is not well understood, there are certain risk factors that the literature suggests contribute to the development of the disease, such as being of African descent, older maternal age, having multiple pregnancies, and poor nourishment (4–6). Unfortunately, PPCM is a global health concern with few studies that provide population-based research to explain how PPCM affects communities across the world. However, a few studies have stressed the strong association between African ancestry and PPCM. Black women have a higher relative risk of PPCM compared to their white counterparts, with a study showing a 15.7-fold higher incidence of PPCM among African American women than non-African Americans (7). With access to health care being a significant factor preventing care, African American women appear to have a poorer prognosis compared to women of other races (7–10). Therefore, the literature recommends that early recognition and diagnosis of PPCM is key to increasing the chance of survival and recovery (11–14).

As a result of these recommendations, U.S. cardiologist, Dr. James Fett developed and validated a self-test instrument to aid women to distinguish the signs and symptoms of PPCM from that of a normal pregnancy. Dr. Fett identified the six common signs and symptoms of PPCM as: 1) orthopnea, 2) dyspnea, 3) unexplained cough, 4) swelling, 5) excessive weight gain during last month of pregnancy, and 6) palpitations; each of these were quantified with a simple scale of 0, 1, and 2 to facilitate self-scoring (15). In the study, the instrument was tested retrospectively among a sample of 47 mothers who were previously diagnosed with PPCM and a control group of 10 non-PPCM mothers. Based on the results, it was recommended that patients with a score greater than 4 should have their providers conduct a blood B-type natriuretic peptide test and an echocardiogram to aid in diagnosis (15). The Fett self-test is readily available and easy to understand; however, when accounting for patients whose primary language is not English, a simple direct translation may obscure the test’s meaning (16).

Translating health screening tools is essential for expanding access to resources and alleviate health disparities cross-culturally. However, direct translations of screening instruments are insufficient to ensure patient understanding. Previous models of translation merely focused on direct and forward translations from the original to the target language (16). However, the literature demonstrates that the lack of adaptation does not allow for patients or evaluators to apply the measure in the local context, while ensuring the original meaning of the instrument is accurate.

In this study, we aimed to adapt the previously validated Fett PPCM self-test for use among a Haitian Creole speaking population. The translation and adaptation of health screening instruments reflect an intricate process that necessitates a systematic approach. Studies have shown that end-users’ opinions of translated health screening tools are often not sought (16). Therefore, this study sought the perspective of both health professionals and community advisory board members to culturally adapt the Fett PPCM measure through the use of focus groups and cognitive interviews.

Methods

Study Design

Data collection took place between May 2019 and July 2019. A qualitative approach was most appropriate for the scope of this study, employing focus groups and cognitive interviews. Each of the systematic steps employed in this process was used to inform the next to preserve the integrity of the adaptations. To begin the study, we conducted a direct translation of the Fett PPCM Self-Test from English to Haitian Creole while making minor adaptations to account for cultural nuances. The original Fett version was translated by two Haitian Creole speakers who were fluent in both languages and verified by a third translator to ensure the most accurate direct translation. A qualitative approach was then employed through the use of focus groups and cognitive interviews. We began the study by conducting a total of four focus groups (n=29) with medical providers in Haiti (Table 1). Two of the focus groups consisted of nurses and midwives, while the other two comprised of obstetricians, primary care physicians, and pediatricians. We began the series of focus groups with the nurses and midwives. As they represent the supporting staff who engage with the Creole speaking patients more frequently, we believed that it was best to conduct the first assessment of the translation and adaptation with them.

Table 1.

Demographic of providers from Focus Group

Group	No. of Participants	Sex	Specialties
1	7	7 Female, 0 Male	5 Nurses, 2 Midwife/Nurses
2	8	3 Female, 5 Male	2 OB/GYN, 6 Primary Care Physicians
3	7	7 Female, 0 Male	5 Nurses, 2 Midwife/Nurses
4	7	4 Female, 3 Male	1 OB/GYN, 5 Primary Care Physicians, 1 Pediatrician

The measure was further refined based on feedback gathered after each subsequent focus group, resulting in a 6-item, culturally adapted screening instrument that would be administered by an auxiliary health care professional or community health care worker. The cognitive interviews were then conducted with the adapted screening tool among 16 Community Advisory Board (CAB) members (Table 2). Upon completion of the focus group or cognitive interview, each participant was offered refreshments equivalent to approximately 175 Haitian gourdes (~ $2.50 U.S.), for their participation in the study.

Table 2.

Demographic of Community Advisory Board (CAB) Member Participants

Participant No.	Gender	Number of Years Working at the Center in the Area of Maternal Health
1	Female	9
2	Female	6
3	Female	9
4	Female	9
5	Female	9
6	Female	9
7	Female	12
8	Female	9
9	Female	9
10	Female	2
11	Female	¼
12	Female	8
13	Male	1
14	Male	5
15	Female	9
16	Male	9

Recruitment

Study participants were recruited from the Groupe Haitien d’Etudes du Sarcome de Kaposi et des Infections Opportunistes (GHESKIO) center in Port Au Prince, Haiti. The physicians and nurses from various departments within the center were contacted and invited to participate in the study focus groups. We also worked with the Director of the GHESKIO Community Advisory Board (CAB) to recruit participants for the cognitive interviews. The GHESKIO CAB includes members from all socio-demographic levels and opinion leaders from all sectors of the population. It was established in 1998 as a link between GHESKIO and the local community; its members regularly serve as advisers on research protocols. Informed consent was obtained from participants who expressed an interest in participating in the study.

Initial Direct Translation

In the direct translation of the measure, we added a three-sentence set of instructions for the personnel administering it to be read to the patient. It stated,

‘I will ask you questions about some signs and symptoms that pregnant and postpartum women usually have. For each question, tell me to what extent you have felt these things during pregnancy or in the first 6 months after giving birth. If you have never experienced them, tell me; also, if you have any of these symptoms, tell me to what extent you feel them.’

It was translated into Haitian Creole as follows:

‘Mwen pral poze w kèk kesyon sou kèk siy ak sentòm moun ansent ak moun ki fèk fin akouche konn genyen. Pou chak kesyon, di m jiska ki pwen ou konn santi bagay sa yo pandan ou ansent oubyen nan 6 premye mwa ou finn akouche. Si w pa janm fè eksperyans sa yo, di m ; tankou tou si ou konn santi kèk nan sentòm sa yo, di m jiska ki pwen ou konn santi yo.’

Focus Groups

Focus groups help with capturing a great deal of information and encourage conversations among participants allowing facilitators to gauge the perspective of multiple people (17).

With the use of the original Fett self-test and initial direct translation, four focus groups were conducted with health professionals to gather their perspective and to begin refining and adapting the initial translation (Table 3). The feedback from health care professionals helped to validate each instrument item. The focus groups were facilitated by two professionals fluent in Haitian Creole. One facilitator led the group discussion with the use of a flip chart, while the second facilitator noted significant comments and suggestions that emerged during the session. Each participant was provided with a physical copy of the original English Fett self-test as a reference, while the initial translations were presented on the flip chart and used to guide the focus groups. While participants had a good understanding of English, some were significantly more fluent than others. Therefore, if any confusion arose regarding the terminology during the focus group, participants were able to talk through it amongst themselves with very limited intervention from the facilitators. Upon completion of each focus group session, the facilitators would review each of the instrument items with the group and ask for a final consensus on the best adaptation based on the suggestions discussed. The facilitators would then move forward with the changes only if the majority of the participants agreed. The final adaptations made by the end of each focus group was based on the sole guideline that it retained the context of the original measure and was simple enough for the target audience to grasp.

Table 3.

Self-Test Instrument Direct Translation and Final Adaptation

Original Fett Version	Initial Haitian Creole Direct Translation	Final Adaptation Haitian Creole	Final Adaptation English Translation
Orthopnea (difficulty breathing when lying flat): (a) None (b) Need to elevate head (c) Need to elevate 45 degrees or more	Difikilte pou respire lè w kouche plat (Orthopnée): (a) Okenn difikilte (b) Bezwen pou leve tèt mwen tou piti (c) Bezwen pou leve tèt mwen 45 degre oswa pi wo	Ou konn gen pwoblem pou w pran souf lè w kouche tèt ba? Si non, chwazi opsyon a. Si wi, mande l «konbyen zòrye ou ta bezwen mete anba tèt ou pou w santi w alez?» Epi chwazi repons la. (a) Pa konn gen okenn pwoblem pou ou pran souf lè w kouche tèt ba. (b) Bezwen pou ou mete yon lòt zòrye. (c) Bezwen pou ou mete plis ke 2 zòrye.	Do you have problems breathing when laying down with your head low? If no, choose Option A. If yes, ask them ‘how many pillows would you need to put underneath your head to feel comfortable?’ Choose the response.  (a) Don’t have any problem breathing when laying down with my head low. (b) Need to add an additional pillow. (c)Need to add more that 2 pillows.
Dyspnea (shortness of breath on exertion): (a) None (b) Climbing 8 or more steps (c) Walking on level	Souf kout lè w fè efò (Dyspnée): (a) Pa ditou (b) Lè w monte 8 mach eskalye oswa plis (c) Lè w mache sou menm nivo a	Ou konn santi souf kout oubyen souf anlè lè w fè efo (santi w bouke)? Si non, chwazi opsyon a. Si wi, mande l «nan ki moman?» Site opsyon b ak c, epi chawzi pi bon repons la. (a) Pa konn gen okenn pwoblem souf kout oswa souf anlè oswa m pa konn bouke lè m fè efo. (b) Lè w monte nenpòt ti mon oswa mach eskalye. (c) Lè w mache nòmalman.	Do you feel shortness of breath or labored breathing on exertion (feeling tired)? If no, choose Option A. If yes, ask them ‘in which instance?’ State Option B and C, and choose the best response. (a) Don’t have any shortness of breath, labored breathing, or tiredness on exertion issues. (b) When you walk up any small incline or stairs. (c) When you walk normally.
Unexplained cough: (a) None (b) At night (c) Day and night	Tous san rezon: (a) Pa ditou (b) Lan nwit sèlman (c) Lajounen tankou lannwit	Wap touse? Si non, chwazi opsyon a. Si wi, mande l «kilè ou konn ap touse? oswa ‘nan ki moman ou konn ap touse?» Site opsyon b ak c, epi chawzi pi bon repons la. (a) Non. M pap touse. (b) Lannwit sèlman. (c) Lajounen tankou lannwit.	Are you coughing? If no, choose Option A. If yes, ask them ‘when do you cough?’ or ‘in which instance?’ State Option B and C, and choose the best response. (a) No. I’m not coughing. (b) Only at night. (c) Day and night.
Swelling (pitting edema) lower extremities: (a) None (b) Below knee (c) Above and below knee	Pye ak janm enfle (enflamasyon) (a) Pa ditou (b) Anba jenou (c) Anwo ak anba jenou	Ou konn gen pye ak janm anfle? Si non, chwazi opsyon a. Si wi, mande patisipan pou li montre w kote ki anfle. Epi chawzi pi bon repons la. (a) Pa konn gen pye ak janm anfle. (b) Nan pye. (c) Tout janm nan.	Do you sometimes have swollen legs and thighs? If no, choose Option A. If yes, ask the patient to show you where is swollen. Choose the best response. (a) Don’t have swollen legs and thighs. (b) In the leg. (c) The entire thigh and leg.
Excessive weight gain during last month of pregnancy: (a) Under 2 pounds per week (b) 2 to 4 pounds per week (c) Over 4 pounds per week	Gwosi twòp pandan dènye mwa gwosès la: (a) Mwens pase 2 liv pa semèn (b) Ant 2 a 4 liv pa semèn (c) Plis pase 4 liv pa semèn	Ou santi w gwosi twòp pandan dènye mwa gwosès la? Site tout chwa yo, epi chawzi pi bon repons la. Verifye pwa pasyan an avèk medsen oswa enfimyè. (a) Ou santi w fè yon ti gwosi. (b) Ou santi w fè yon bon gwosi. (c) Ou santi w gwosi anpli anpli.	Do you feel like you have gained too much weight during the last month of the pregnancy?  State all of the options and choose the best response. Verify the patient’s weight with the doctor or nurse. (a) You feel you have gained a little bit of weight. (b) You feel you have gained a good amount of weight. (c) You feel you have gained a lot, a lot of weight.
Palpitations (sensation of irregular heart beats): (a) None (b) When lying down at night (c) Day and night, any position	Sansasyon batman kè iregilye (Palpitasyon): (a) Okenn (b) Lè w kouche lan nwit sèlman (c) Lajounen tankou lannwit, nan nenpòt ki pozisyon	Ou konn santi sansasyon batman kè (kè w ap bat fò oswa kè w ap pile)? Si non, chwazi opsyon a. Si wi, mande l « kilè ou konn gen sansayon sa ? » Epi chawzi pi bon repons la. (a) Pa konn santi sa. (b) Lè w kouche lannwit sèlman. (c) Lajounen tankou lannwit, nan nenpòt ki pozisyon.	Do you feel the sensation of your heart beating forcefully (your heart is beating rapidly or your heart is pounding)? If no, choose Option A.  If yes, ask them ‘when do you have that sensation’? Choose the best response.  (a) Don’t feel that. (b) Only when you lay down at night. (c) Day and night, in any position.

Cognitive Interviews

Following the adaptations gathered from the feedback of healthcare providers during the focus groups, the cognitive interviewing process allowed for a form of pre-testing among community board members whose sociodemographic characteristics were similar to that of the target population (18). We worked with the Director of the GHESKIO Community Advisory Board to recruit CAB members who interact with the community; participants included women of child-bearing age who could relate to the context of the self-test instrument. During the cognitive interviews, individuals were able to help with rephrasing and identifying problematic questions or statements; the process gave them the opportunity to restate questions in their own words to ensure they understood the question as intended (19). It is essential that changes made to the assessment tool accurately measure symptoms in the same manner as the original instrument.

Ethics Approvals

The study protocol was approved by the GHESKIO Ethics Committee, Cornell University Institutional Review Board (IRB), and the Florida International University (FIU) Institutional Review Board (IRB).

Results

Focus Groups

The primary concern among the providers related to the responses that required patients to count or measure. A large proportion of the target population have little to no education, therefore providers stressed that any questions involving interpretation of large numbers or unknown concepts would make it difficult for the target population to grasp, potentially resulting in inaccurate responses.

The first item of the instrument assessing orthopnea had an option describing the need to elevate the head 45 degrees or more. This response option assumes that the patient has the ability to measure what 45 degrees means. Therefore, providers agreed to use the number of pillows as an equivalent concept. The number of pillows mentioned was no more than three, which is simple enough that patients could differentiate it with ease. Providers discussed the importance of emphasizing the ‘need’ in this first item, so that patients understood that though they may not have necessarily carried out this action, they felt a need to do so in order to relieve their difficulty breathing when lying flat. For example, patients may feel a need to use three pillows to alleviate their difficulty breathing, however, they may not have had three pillows to do so.

The second item assessing dyspnea, includes climbing 8 or more steps. Providers noted that many patients may not normally walk up a flight of stairs, therefore patients may be unable to conceptualize what is comparable to climbing 8 steps. Thus, providers agreed to use a small incline as an equivalence, as they are more commonplace in the city and the patients were more likely to encounter them. However, they also agreed that it was important to still include climbing stairs, but to remove ‘8 steps’.

The third item asked if the patient had unexplained coughing; however, providers suggested that based on their experience with patients, they would more often than not respond to this question with a ‘yes.’ Thus, it was key to follow up the question by asking: ‘when do you experience the coughing?’ Based on their reply, interviewers or providers could determine which of the three response options was most appropriate (Table 3).

The fourth item, swelling in the extremities, was considered challenging because patients might find it difficult to determine the specific location of swelling in their lower extremities; the options included: 1) None, b) below the knee, and c) above and below the knee. However, when directly translated into Haitian Creole, providers found it difficult to understand and thought it would likely cause confusion. It was suggested that the question specify ‘nan pye’ (literally: in the foot) for below the knee and ‘tout janm nan’ (literally: the entire leg) for above and below the knee. Providers thought it would be important to have the individual administering the instrument have the patient point to the specific area of their leg that was swollen. The provider would then select the appropriate response based on where on the leg the patient had indicated.

The fifth item assesses excessive weight gain during the last month of pregnancy. The options were: a) under 2 pounds per week, b) 2 to 4 pounds per week, and c) over 4 pounds per week. However, providers also found this question to be problematic. Access to resources like a scale and forming the habit of regularly weighing themselves is not commonplace among the population in Haiti. Therefore, providers believed that though patients’ responses may be inaccurate, asking them this question was still important to evaluate their perspective of their weight gain during the pregnancy. Providers decided to adapt the response choice to state, ‘a) ou santi w fè yon ti gwosi, b) ou santi w fè yon bon gwosi, and c) ou santi w gwosi anpil, anpil’ (Table 3). The literal translations of the adaptations were: ‘a) you feel you have gained a little bit of weight, b) you feel you have gained a good amount of weight, c) you feel you have gained ‘a lot, a lot of weight’ (Table 3). However, the personnel administering the questionnaire would either follow up with the patient’s provider or access their documented weight on the electronic medical record (EMR) to note the proper response on the questionnaire.

The last item of the instrument assesses palpitations (sensation of irregular heartbeats). Providers suggested removing the word ‘irregular’ in the Haitian Creole translation because patients would not be familiar with the term; however, they would be familiar with ‘palpitation’ and ‘heartbeat.’ Therefore, it would be important to include certain probes to aid patients in describing their perception of irregular heartbeats, for instance, the use of a familiar onomatopoeia such as ‘bidip bidip’ to describe to patients the sensation of rapid heartbeat. Providers also discussed accounting for the phrases that patients often use to describe the sensations they often feel in relation to their heartbeat (Table 3).

We observed that the nurses and midwives were highly focused on basing their proposed adaptations on the realities and personal accounts their patients had shared with them. In contrast, although the doctors were sympathetic to their patients’ experiences, their concern was also to ensure that the integrity of the measure’s clinical value was preserved. Since the four focus groups occurred in an alternating pattern of: nurses/midwives, physicians, nurses/midwives, and physicians, we were able to build on the changes from each type of provider and safeguard the original intent of the test.

Cognitive Interviews

During the cognitive interviews, the CAB participants reported that the instructions section was specific and direct enough to convey the purpose of the questionnaire. They also noted that all the questions were comprehensive and sufficiently clear for patients to understand. However, to ensure consistency with Item No. 4, about the location of swelling, they recommended adding a diagram as a visual aid that specifically defines what constitutes the ‘janm’ and ‘pye,’ as described previously under Item 4. There was no significant feedback obtained from the CAB participants aside from the suggestion of utilizing a visual aid for Item 4.

Discussion

The findings from the study suggest that a direct translation of the Fett Self-Test measure was inappropriate for a primarily Haitian Creole speaking population. Various studies on the cultural adaptation of health instruments have stressed a similar perspective regarding the inappropriateness of relying solely on direct translations (16, 20–24). The processes described by these studies were similar to the current study as they included an initial forward and back-translation of their respective instrument by individuals fluent in the native language. This translation was then followed by a review by another individual or a panel to help assure the validity of the translation and adaptation. The Haitian Creole PHQ-9 study on the reliability and validity of the instrument for instance, followed this approach and tested the tool among a sample of 1,080 participants in Haiti (23). It was found to be a valid measure for use in assessing depression among participants. Another study conducted among a Lebanese speaking population, translated and adapted the Adolescent Pediatric Pain Tool (APPT) following the schema of forward and back-translations, expert panel, pre-testing and cognitive interviewing (22). Their process yielded an effective tool that is both culturally sensitive and reliable.

Though we gathered the focus groups based on profession, it is important to note that integrating each focus group with doctors, nurses, and midwives may also have been beneficial. It would have allowed the focus group discussions to have a wider set of perspectives and stimulate conversations that may not have emerged by keeping participants grouped based on profession. However, ‘power structure’ could be a concern when blending the focus groups; nurses and midwives may feel less expressive or intimidated to voice their opinions with physicians present (25–27).

To account for the low literacy rate of the target population, tailoring the self-test measure into an interviewer-administered instrument was most appropriate for the population. Therefore, the adapted measure included an instruction section followed by the six common signs and symptoms of PPCM in questionnaire form. The instrument provides short probes for each of the items to account for the variation in patient responses. The most challenging obstacle of the adaptation process was ensuring that the original meaning of the self-test was not lost throughout the focus group and cognitive interview adaptations.

Limitations

This study had several limitations. Pregnant and postpartum women were not utilized to evaluate the Haitian Creole adaptation of the instrument. This decision was purposefully taken for ethical reasons, to avoid alarming the women about potential problems without having the infrastructure and linkage in place to provide follow-up care if the women were to present with the signs and symptoms of PPCM. The participants in this study were all recruited from one site, therefore the reach and perspectives were limited to those who were associated with GHESKIO. Providers and community members outside of GHESKIO may not have the same level of formal training, therefore their interpretations of the instrument’s items may be different from the interpretations of GHESKIO staff. A supplemental document, that provides a brief description and objective of each question could be used in conjunction with the measure to ensure that interviewers or providers understand what each question is assessing when administering the instrument with their patients.

Conclusion

The strength of this study is that it provides an instrument for a primarily Haitian Creole speaking population to help understand the signs and symptoms of PPCM during late pregnancy and the early postpartum period. It was adapted based on feedback from medical professionals and individuals who closely identify with the community. No similar screening tool currently exists. Though some of the instrument’s components were easily translated with minor adaptations, several benefited from a multi-stage adaptation and refinement process. This study has resulted in a Haitian Creole version of the Fett Self-Test instrument, which can be administered by auxiliary health care personnel and community health care workers to assess the need for further evaluation for PPCM among patients. The availability of this measure will allow for use in clinical as well as non-clinical settings and potentially in future preventative studies on peripartum cardiomyopathy in Haitian Creole communities. It is recommended that further research be conducted to validate the adapted instrument.