We read with great interest the article by Fischer et al. (1) that was recently published in Diabetes Care. Although we congratulate the authors for their encouraging findings on the impressive role of text messaging support for weight loss management in prediabetes, their results raise several important concerns.
It is very surprising that weight loss was observed only among Spanish speakers but not in the English-speaking group (1). Do the authors feel confident about the internal validity of their findings (i.e., non–quasi-experiment design), considering the potential diversities in educational status, income, type of cell phones (touch tone, large-screen displays), socioeconomic background, and type of personality (motivation, incorporation of family and friends into the weight loss program) among their trial arms of English and Spanish native speakers?
Is the presence of other comorbidities a contributing factor? Although terminal diseases with life expectancy <12 months and diabetes were excluded, information about other chronic non–life-threatening conditions, namely asthma, hypertension, and cardiovascular diseases, is missing. These comorbidities may diminish the global quality of adherence to text message by resulting in “compliance fatigue.”
Are different baseline characteristics also contributing factors? A higher ratio of women and higher mean baseline HbA1c (6.0 vs. 5.8 mmol/mol) could prompt more serious actions to reduce weight in the Spanish group due to a greater level of self-care. Conversely, English-speaking subjects had higher average systolic blood pressure and weighed significantly more than the Spanish speakers (1).
The content and protocol of the English and Spanish messages were not disclosed. What were the exact lines and types of messages used in each allocation group? Were the Spanish messages direct translations from the original English (or vice versa), or were they slightly different in terminology? Could it be that observed weight loss specific to the Spanish speakers is a result of the messages being better tailored to the Latino community, as the authors state (1)?
Fischer et al. (1) cited a small selection pool, single safety net health care system, and higher allocation of Spanish patients for this language-based impact and were unaware of any description in the literature concerning the variable impact by language of text messaging support (1).
A study of Spanish and English speakers with diabetes found enhanced delivery of patient education and self-care by automated telephone calls among Spanish speakers (64% vs. 36%, respectively) (2). This significant difference (P < 0.0001) was interesting because Spanish language messages were not customized based on disparities in culture or vernacular among Hispanic groups of different nationalities (2). In depressed people who received text messages for support, only Spanish-speaking (compared with English-speaking) patients mentioned that receiving messages made them feel as if someone cared for them (3). There are other descriptions on the Spanish-specific efficacy of text messaging support in the literature (4). These observations suggest a systematic, not incidental, contribution of language to text message supportive care.
Understanding the origins of this language-specific impact is essential to determine the best-suited modalities (automated call or text, scheduled e-mail, web-based user interface, etc.) to certain ethnic/language groups for state-of-the-art patient-provider communication beyond the horizon of clinical practice.
Article Information
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
