A descriptive study of potential participant preferences for the design of an incentivised weight loss programme for people with type 2 diabetes mellitus attending a public hospital in Lima, Peru [version 2; peer review: 2 approved]

Background: Weight loss is important for the control of type 2 diabetes mellitus but is difficult to achieve and sustain. Programmes employing financial incentives have been successful in areas such as smoking cessation. However, the optimum design for an incentivised programme for weight loss is undetermined, and may depend on social, cultural and demographic factors. Methods: An original questionnaire was designed whose items addressed respondent personal and health characteristics, and preferences for a hypothetical incentivised weight loss programme. One hundred people with type 2 diabetes mellitus were recruited to complete the questionnaire from the endocrinology clinic of a public hospital in Lima, Peru. A descriptive analysis of responses was performed. Results: Ninety-five percent of subjects who had previously attempted to lose weight had found this either 'difficult' or 'very difficult'. Eightyfive percent of subjects would participate in an incentivised weight loss programme. Median suggested incentive for 1 kg weight loss every 2 weeks over 9 months was PEN 100 (~USD $30). Cash was preferred by 70% as payment method. Only 56% of subjects would Open Peer Review

a recent study testing a behavioural weight loss intervention for Latinos in the United States concluded that companionship for physical activity appears to support weight loss 20 . It is possible that the success of an incentivised weight loss intervention might be optimised by accounting for the social and cultural characteristics of its target population, and by incorporating beneficial social support by design.
Healthcare in Peru is funded publicly and privately: approximately 30% in the lowest socioeconomic stratum is covered by public health insurance (SIS); a further 25% are covered by social security (EsSalud) linked to their employment; 2% have private health insurance; and 38% have no health insurance. Separate military, police and other systems account for the remainder. Insulin, metformin and glyburide are available through SIS and EsSalud, while many further agents are available privately. Glucose testing strips are available for insulin-dependent diabetes through SIS and EsSalud, or through private insurance. There is no national strategy for diabetes care which integrates medical therapy with promotion of exercise or healthy diet 21 .
The aim of this study was to investigate the optimal design for an incentivised weight loss programme which is planned for people with T2DM in Lima, Peru (to be funded by research grants). The objectives of the study were: to determine the acceptability of financial incentives for weight loss among type 2 diabetics in Lima, Peru; and to determine the optimal amount and delivery method for such an incentive.

Design and data collection
We performed a cross-sectional exploratory study using an original questionnaire. Interviews were conducted and data recorded by JP.

Questionnaire development and design
The questionnaire was developed by the authors and not validated separately. It consisted of 82 items (see Supplementary Material for the instrument in original Spanish and translation) addressing socio-economic circumstances, health characteristics and preferences relating to a proposed incentivised weight loss programme. Items relating to the programme included a suggested incentive amount and identifying a threshold incentive amount.
Two methods were employed to identify threshold incentive amounts for participation in a weight loss reduction programme: direct questioning and fixed-increment questioning (Supplementary Table 1 and Supplementary Table 2).
For the first method, a hypothetical situation was explained to the participant, which consisted of inviting them to participate in a 9-month programme whose purpose was to pay a monetary incentive only if they lost 1 kilogram every two weeks, and that we were interested in knowing the exact amount of money that would motivate them to lose that kilogram. For the second method, amounts of money from 0 PEN to 250 PEN in fixed increments of 50 PEN were specified and the participant

Introduction
Weight control is critical for both prevention and treatment of type 2 diabetes mellitus (T2DM) 1-3 . Self-management programmes for people with T2DM commonly include the promotion of lifestyle changes, such as dietary modifications and increasing physical activity, to reduce weight 4-6 . However, sustained weight loss is a challenge to both patients and providers: failure to sustain weight loss in formal diet programmes varies between 21-54%, and many people fail repeatedly 7-9 .
A major challenge in any lifestyle intervention programme is the willingness to join, and sustain, participation. Better understanding of what motivates people to engage with such programmes is therefore fundamental to their design 10 . Financial incentives have emerged as strategies which can initiate and sustain positive health behaviours during the incentive period and beyond. Sustained changes have been achieved through incentivization in the field of smoking cessation, but trials of financial incentives have previously failed to achieve sustained weight loss 11-14 .
Social and cultural factors influence participants' engagement with weight control: for example, among adolescents, weight loss attempts were more frequent in Latinos than Whites or African Americans; desire to lose weight is more common in Latino females than males; and intention to lose weight is related to the number of social contacts trying to lose weight [15][16][17] . Successful completion of both short-and long-term weight loss programmes has been associated with age, ethnicity, family structure, educational level and employment 18,19 . Additionally, • A revision of references in the introduction in accordance with peer review recommendations.
• On request of both peer reviewers, a paragraph added to introduction summarising diabetic care in Peru.
• Revision of statement of aims and objectives to improve clarity.
• Structure of methods revised for clarity.
• Provided main questionnaire in English and Spanish (Supplementary File 1 and Supplementary File 2) • Substantial modification to discussion section on recommendations from both peer reviewers: -Further discussion of the finding that more participants would participate in unincentivised than incentivised programmes -Further discussion of the reservations which some participants expressed regarding the morality of financial incentives -Elaboration of limitations of study, in particular the absence of anthropomorphic and laboratory data.
-Discussion of the problem of weight loss maintenance in the proposed 9-month incentivised programme Numerical errors in Table 3 corrected.
was asked whether each of these amounts would motivate them to lose 1 kilogram over two weeks.
Participants were also asked about their willingness to participate in a hypothetical 'deposit-contract' programme in which they would be required to deposit a certain amount of money in a saving account and such amount would be doubled if they lost 1 kilogram over a two-week period, but would lose the deposited amount if they failed to reach the weight loss goal.
Finally, participants were asked if they would be willing to share the money won in a weight loss programme with a coparticipant, defined as a relative or friend selected by the participant to support their efforts to lose weight, their preferred co-participant, and the proportion of the incentive that the participant would be willing to share with this co-participant.

Participants
Patients were recruited by convenience sampling from the Hospital Nacional Arzobispo Loayza, a public tertiary hospital serving mostly low-income people from Lima, the capital city of Peru, whose endocrinology department provides over 2500 outpatient appointments annually to patients with T2DM 22 .
Inclusion criteria were age ≥18 years and self-reported diagnosis of T2DM. Incapacity to provide written informed consent was the only exclusion criterion. As patients were attending an endocrinology clinic it was not considered necessary to verify their self-reported T2DM status independently, while the research team did not have access to participants' medical records. Due to the exploratory nature of the study, only 100 subjects were invited to participate. Participants were recruited in the waiting room of the Endocrinology Department during April 2016.

Data analysis
A descriptive analysis of questionnaire items was undertaken, employing 95% confidence intervals for selected items whose measurement was considered particularly important. For nonparametric continuous variables, a bootstrap confidence interval of the median was attempted. Hypothesis testing was not performed due to the large number of possible comparisons relative to the sample size and the consequently elevated risk of type 1 error. Statistical analysis was performed using R version 3.4.3 23 .

Ethics
This study was approved by the Institutional Review Boards of the Universidad Peruana Cayetano Heredia (SIDISI 64789) and the Hospital Nacional Arzobispo Loayza (Expediente 04974-2015), in Lima, Peru. Written informed consent for participation was obtained from all subjects.

Results
One hundred people with T2DM participated in the study. Two subjects did not respond to questions relating to incentives; the data were otherwise complete. Demographic and socioeconomic characteristics are presented in Table 1. Health-related responses are presented in Table 2. Measures previously taken to improve health are presented in Table 3.
Ninety-eight subjects (98%) responded to questions about financial incentives. Ninety-two subjects (94%; 95% CI 87 -97%) responded that they would participate in an unincentivised weight loss programme. Eighty-three (85%; 95% CI 76 -91%) would participate in a 9-month incentivised weight loss programme. Reasons given for not participating included: insufficient time to attend biweekly follow-up visits; because they thought it would be difficult to avoid "antojitos" (cravings) for 9 months; or because the participant did not think they needed to lose weight. Seventy-eight subjects (78%) answered the question "how much money would motivate you to lose 1 kg every 2 weeks?". Responses were positively skewed with median PEN 100 (≈ USD $30) and range PEN 50 to 500 (≈ USD $15 to 150) ( Figure 1). Bootstrap confidence intervals could not be constructed because all resampled medians = PEN 100 (10,000 simulations).
Subjects were then asked whether they would participate in an incentivised weight loss programme with incentive amounts from PEN 50 to 250 in PEN 50 increments. Six subjects (6%) would not participate for any amount, while 91 (93%) would participate for all amounts. One subject changed from a positive to negative response at the PEN 200 threshold.
Asked about their preferred method of payment, 69 subjects preferred (70%) cash, 24 (25%) deposit into a bank account, 3 (3%) as vouchers and the remainder not responding.
Fifty-five subjects (56%; 95% CI 46 -66%) would participate in a deposit-contract scheme whereby their deposit would be doubled if they succeeded but lost if their failed to lose weight. Ninety-seven (97%) subjects answered a question on preferred deposit amount. Preferred deposit amount was positively skewed with median PEN 20 (≈ USD $6) and range PEN 0 to 50 (≈ USD $0 to 15) ( Figure 2). Again, equality of all resampled median precluded construction of bootstrap confidence intervals.
Subjects were then asked whether they would participate in a deposit-contract scheme with deposit amount in increments between PEN 25 -250. Forty-three subjects would participate with any deposit amount (43%); 32 would not participate with any deposit amount (32%); and 22 identified a threshold deposit amount for participation (22%). Among subjects who identified a threshold deposit amount above which they would not participate, the maximum acceptable amount was positively skewed with median PEN 25 (range PEN 25 to 100).
Regardless of their answers to the previous questions, subjects were also asked for their views of participating in such a program. Out of the 73 who responded, 14 (19%) considered that it was not good to receive money for taking care of their own health, with one saying that this would be "like selling yourself", since people should lose weight for their own sake and not for money. Sixteen (22%) said it was a good idea and were even excited at the prospect of participating in the program. Six (8%) found it amusing that such a program was even possible, and five (7%) were concerned that such a program will achieve only short-term results that would not be sustained after the program ended. Other answers revolved about the doubts they had about the program, or they did not understand the idea behind receiving money, that it was a good idea for "poor" people but not for everybody or that it might not work since not all diabetics needed to lose weight.

Discussion
This pilot study aimed to characterise people with T2DM attending a public hospital in Lima, Peru, and their preferred amount and delivery method for a financial incentive to be used in a future incentivised weight loss programme.
The proportion of participants who would participate in an incentivised weight loss programme was high (85%) but the proportion who would participate in an unincentivised programme was even higher (94%). A similar pattern was observed in a mixed-methods study of acceptability of incentives for a weight loss maintenance programme, in which 93.9% supported the programme generally but only 77% supported cash incentives 24 . The finding may indicate that in our sample weight loss as a goal was a more powerful motivator than the financial incentive. However, this does not negate the potential utility of incentives, which might contribute to participant retention and sustained weight loss achievement in addition to recruitment. It is possible another group exists, but was not accessed in this study, who do not wish to lose weight but could be motivated by financial incentives to do so. Such a group might be the most appropriate target for an incentivised weight loss intervention, but its access could constitute a significant challenge. That fewer respondents would participate in an incentivised than in an unincentivised programme may also be due to unacceptability of financial incentives in this population for moral reasons: 14 thought it wrong to accept money in exchange for taking care of your own health, with one describing this as "selling yourself". Similar concerns were expressed in focus groups in a recent study, in which discussion of financial incentives conveyed "distrust and indignation", where the idea was reiterated that improved health should be sufficiently motivating for weight loss 24 . These comments may represent a significant cultural attitude towards financial incentives for health which could constitute a barrier to their success, and which deserve further attention. However, such reservations may not necessarily preclude participation: 85% of subjects nonetheless indicated that they would participate in an incentivised intervention, and a recent systematic review found that participation may actually be increased by financial incentives for weight loss 14 .
Median suggested incentive amount was PEN 100. Based on a national disposable income of USD $175.7bn 25 and population of 30,565,431 in 2013 26 , a maximum reward of PEN 100 every 2 weeks for 9 months would represent 10% of personal disposable income (PDI). Previous interventions have employed a broad range of incentive sizes (from 0.2% to 10.2% of PDI 13 ), and experimental evidence suggests that insufficient incentives may paradoxically produce less motivation than no incentive at all 27 . The suggested amount therefore appears adequate and appropriate for an intervention in this setting.
The fixed-increment questioning method to identify a suitable incentive amount (asking whether the participant would accept amounts of increasing PEN 50 increments) was not successful. Sixty-two percent of participants in a previous study felt that financial incentives undermined individual responsibility for health 28 , and participants may have been reluctant to engage with these questions to avoid weighing a moral position against financial advantage.
Fewer respondents would participate in a deposit-contract scheme, which concurs with previous findings 24 . Because such schemes weigh a certain short-term price against a possible longterm advantage, they fail to take advantage of the established health economic principle that individuals overvalue present relative to future costs 29 . In contrast, an approach described as asymmetric paternalism, which aims to assist individuals with health-improving behaviours without limiting freedom 30 , might produce in an intervention in which individuals commit to future behaviours without present costs, such as receiving up-front an incentive which would be returned or doubled depending on achievement of a future weight goal. Cash or bank transfer were generally preferred over vouchers. This is in accordance with the finding that rewards are more motivating when separated from larger payments, such as household shopping (in the case of vouchers) and insurance premiums (in the case of discounts) 31 .
Our findings show that most participants had found it challenging to adopt health-improving behaviours. In particular,   42% of participants had previously attempted to lose weight but 95% found this "difficult" or "very difficult", suggesting that people with failed previous weight loss attempts will constitute a substantial subgroup of this population. The question of what makes behavioural change difficult has been addressed by Kelly & Barker, who note the mistakes which policy-makers commonly make in understanding the drivers of behaviour 32 . One of these mistakes is the economic utility theory which presumes that individuals make rational choices to maximise gain and minimise loss. The theory behind the use of financial incentives is essentially an extension of this. However, health behaviours are frequently automatic responses to social and environmental cues, not subject to particular conscious reflection, and often in spite of adequate understanding of health implications 33-35 . These findings inform interventions that target 'choice architecture', comprising the "interaction between individual human agency and both the immediate and broader environment that make up the social structure" 36 . Financial incentives are much more likely to achieve persisting behavioural change in synergy with such interventions.
Asked about who they would choose to help them to lose weight, most selected a family member. In prospective studies, family support was associated with reduced HbA1c in males, but increased HbA1c in females. Informal support seeking is often different in males and females. Females seek and receive more support from friends and extended family, while males often seek and receive more support from their spouse 37 .
Other studies found that seeing friends more frequently, having a well-functioning social network and a sense of good social support from the social network was associated with higher patient activation levels, less diabetes-related emotional distress and more health-promoting self-management behaviours among patients with T2DM. When providers felt more emotionally engaged, their support exerted a large, positive effect on their well-being, as well as on recipients' well-being 38,39 . These findings imply that the incorporation of social support into an intervention may be crucial for its success, but also that its precise form may need to adapt to the sex (and potentially other characteristics) of the participant.
Five participants additionally raised concerns over the sustainability of weight loss in such a programme, and indeed this obstacle remains to be overcome: thus far, incentivised weight loss programmes have failed to achieve sustained weight loss beyond the incentivised period 13,14 . The Stages of Change model proposes pre-contemplation, contemplation, preparation, action and maintenance stages to behavioural change 40 . To participate in the hypothetical 9-month intervention proposed to participants, they would necessarily have reached the 'preparation' stage but at the conclusion of the intervention might have spent only a maximum of 3 months in 'maintenance', presuming no relapses. Sustained weight loss would therefore not be incentivised for a long period in such an intervention. Indefinite incentive payment is unlikely to appeal to insurers or healthcare providers, nor is it likely to be evaluated due to its likely limited appeal to research funding bodies. Incentives may ultimately prove most valuable in initiating and achieving short to medium term change but will need to be integrated into a multimodal approach for treating T2DM and obesity more generally, including psychological, medical and potentially surgical methods (although these are unlikely to be available to this demographic for some time yet).

Limitations
The sampling approach employed may have exposed the study to participation bias. Most participants were female, middleaged, and had at least completed secondary education. Although most rated their economic status as at least 'fair', almost all had either the most basic or no health insurance at all. Although the prevalence of T2DM is greater in males than females worldwide 41 , the higher proportion in our study may be explained by the fact that females are more likely than males to engage with healthcare seeking behaviours and respond to questionnaires 42,43 .
The study setting in a Peruvian public hospital is likely to have determined participants' socioeconomic profile, which should not be interpreted as representative of people with diabetes in Peru more generally. However, the prevalence of T2DM is inversely proportional to socioeconomic status 44,45 , and therefore the majority of people with T2DM in Peru will fall into the low-income group surveyed in this pilot and targeted by our planned intervention. Higher-educated subjects have previously been found to make more attempts to lose weight 46 , which may imply a greater need for intervention in this low-income group.
Anthropomorphic and laboratory data relating to participants' weight and diabetic control were not recorded and it was therefore not possible to examine whether responses were influenced by these. It is also unknown what proportion of participants were overweight or obese. The possibility exists that although participants reported that had T2DM and were attending an endocrinology clinic they may not have had T2DM, as this was not verified by laboratory testing because the authors did not have access to participants' medical records.
The questionnaire used was original and not previously validated. Important parameters for an incentivised weight loss programme were not explored in our questionnaire. A 'lottery' form for payments, in which successful weight loss would allow entry into a regular lottery for a larger payment (and which is anticipated to be more motivating than direct payments because people tend to over-value small odds of large rewards 31,47 ) was not proposed to participants. Participants were also not asked about their preferred frequency of payment. Higher-frequency payment have been shown to be more effective in the drug-abstinence setting 48 , and the finding that experimental subjects prefer to segregate than to integrate gains has been used to support the argument for direct rewards over insurance premium adjustment 31,49 . These factors are important for the planning of any intervention and the preferences of potential participants should be the subject of future investigation.
Although the difficulty which participants had experienced in adopting health control behaviours was quantified, participants were not asked why each behaviour was difficult. This information could be usefully obtained through qualitative research, and might point to other potential targets for intervention, such as psychological, environmental and social factors.
Although multivariate associations could not be investigated due to insufficiency of sample size and sampling design, the study was not designed to investigate these, but rather to develop an improved understanding of the potential use of incentives in this setting.

Conclusion
The use of direct financial incentives in a future weight loss programme for people with T2DM in Lima, Peru was acceptable to the majority of participants in this study, although some expressed reservations regarding the morality and sustainability of such a programme.

Data availability
Original and translated data files are available on Open Science The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Supplementary material
Supplementary

INTRODUCTION
Number and quality of references 1.
The authors should avoid loading the introduction of references. The introduction 22 references. All of them are necessary? 4 to 5 good quality (primary sources, review or metaanalyses) references is enough. For example: a) Page 3, left column, paragraph 1, line 2: "Weight control is critical for both the prevention and treatment of type 2 diabetes mellitus (T2DM) (1-4)." There are 4 references supporting it. The first reference is a report. The second is appropriate. References 3 and 4 are secondary sources belonging to sections of the American Diabetes Association Guidelines. I suggest looking for a systematic review supporting the importance of weight control in the prevention and treatment of T2D. b) Page 3, left column, paragraph 1, line 2-5." Self-management programs for people with T2DM commonly include the promotion of lifestyle changes, such as dietary modifications and increasing physical activity, to reduce weight (5-7)." There are 3 references but none of them is a primary source. The reference 5 is ADA guidelines and the other 2 are web-pages containing information for patients. I suggest retiring these references (5,6,7) and include primary sources of studies demonstrating the effectiveness of structured weight loss interventions (not only self-management programs) on diabetes management. The most important structured interventions evaluating the effect of lifestyle changes on T2D management are the Look AHEAD study using HbA1c as main outcome and the Why Wait Study whose primary outcome was the weight 1,2 . c) Page 3, left column, paragraph 1, line 5-6. "However, sustained weight loss is a challenge to both patients and providers (8,9)." References 8 and 9 do not mention directly a sustained weight loss as a challenge. Both studies are interviews with physicians. In the reference 8, 14 GP were interviewed in focus groups. Five dilemmas were identified for primary physicians, but none of the 5 referred to difficulties of their patients in losing or maintaining weight loss. Reference 9 was an interview to 19 physicians to understand their challenges when treating social and emotional difficulties in T2D patients, but not the challenge of sustained weight loss to both patients and providers.
It is very laborious as a reviewer to check all references one by one. I strongly recommend that all references be carefully reviewed to establish if they are relevant, useful and what is more important if they support the written statements.

Background
In the introduction, the authors have to create a strong background that explains the reasoning behind why the study goal was built. In the introduction, the authors broadly mention sustainability as one of the possible benefits of adding incentives to a weight loss program. However, a probably greater sustainability of the loss of weight generated by the addition of incentives should not be proposed without bases that sustain it. Page 3, left column, paragraph 2, line 7-9. "Sustained changes have been achieved through in the field of smoking cessation, although this remains a challenge to weight loss interventions." I consider that the comment regarding that sustainability in weight loss interventions is a challenge is very conservative. Two of the references cited deserves attention. In the 2.
reference 15, financial incentives in 66 US veterans with BMI between 30 and 40 and age between 30 and 70 years produced significant weight loss over an 8-month intervention; however, participants regained weight post-intervention. The reference 16 is a systematic review included nine randomized controlled trials of behavioral treatments for obesity and overweight involving the use of financial incentives with reported follow-up of at least 1 year. No study was analyzed on an intention to treat basis, participants were mostly women recruited through media advertisements, mean age ranged from 35.7 to 52.8 years, and mean body mass index from 29.3 to 31.8 kg/m −2 . Results from meta-analysis showed no significant effect of the use of financial incentives on weight loss or maintenance at 12 months and 18 months.
With this background (under-registered in the article), the first important question in the introduction would be: In previous studies, monetary incentives have shown some benefit in patients who are in a weight loss program? If the answer is NO. Is it justified to evaluate it in this population since it has very different characteristics than those evaluated previously, and the results could be different? The same participants question the usefulness of the incentives to increase the willingness to participate since 92% of them reported agreeing to participate in a weight loss program without incentives. Also, 19% considered that. This aspect should be mentioned in the discussion.

The gap of knowledge
In the introduction, the authors must highlight the existing knowledge gap that this study will fill and the relevance of the research question. What is the gap? The need a questionnaire (or the generated information) to determine the acceptance of incentives or to explore the opinion of potential patients regarding the incentives? This gap must be the prelude the research question and the aim. 3.

The aim
It should be brief and should make clear what is the question that your study tries to respond.
The need for a questionnaire to determine the acceptability of monetary incentives and their characteristics within a weight loss plan? Although the introduction was designed to create the need for information on the design of a monetary incentive program for diabetic patients who would join a weight loss program, the first objective mentioned was the characterization of the population. I believe that they should be more direct and go to the main objective. Reconsider: we performed a questionnaire study of potential participants with the aim of defining their demographic, social, cultural and health characteristics… 4.

METHODS
a) I suggest including a sub-heading to explain the structure of the questionnaire and a summary of how was developed. b) Inclusion criteria should be T2D diagnosed by laboratory tests available in the files of the endocrinology clinic and / or the use of antidiabetic drugs, and not only by self-report of the participant c) Considering that the program is aimed at overweight and obese patients, why you did not consider a BMI <25 as an exclusion criterion? The perspective of the problem is different in a person of normal weight with respect to an overweight / obese person. How many of the included participants have a normal weight?

RESULTS
a) The results describe the demographics and socioeconomic characteristics of the T2D patients commonly seen in the endocrinology clinic, considering that was a convenience sample b) Table 2 should include BMI, most recent glucose and HbA1c measurement taken from the hospital files instead of self-reported data c) Table 3 shows that 53 participants reported regular exercise as a health control measure attempted, but when the frequency of difficulty of attempt to regular exercise is reported, the total of participants is 34. I observed the same discrepancy, in the reduction of sugar intake (75/65), to quit alcohol (31/29), to reduce fat intake (77/65). In the case of the increase of vegetable intake (57/57) and weight loss (42/42) no discrepancies between the number of participants reporting the behavior and the total of the three categories of difficulty. Please, correct it if it is a mistake. d) In table 3 methods are included for weight loss control (3 for each participant). Why to avoid appointments and foot care are included? These methods have not a relationship with weight loss.

DISCUSSION
The discussion includes 1. How much the intervention (incentives) represent the proportion of personal disposable income (PDI)? 2. Explanation of the answers of participants about the methods of payment, 3. Comparison with other studies regarding the effect of a helper. I recommend: a) Re-organize the discussion b) Explain the economic utility theory more clearly c) Consider including in the discussion important topics below

Topic 1. Potential implementation
To provide some clues about the implementability, could you include a paragraph mentioning how is the usual care of the diabetic in the hospital and if the endocrinology clinic has a structured program of weight loss? How would monetary incentives be inserted into that program? Who could be the provider of the funds? If you plan to start as a research project?
A Mayo Clinic group proposed a multispecialty outpatient Obesity Treatment Research 1.
Program for weight loss to be implemented in the next 5 years that initially will start with research funds 3 .
Future implementation of a monetary incentive strategy for weight loss must consider different elements. Identification of stakeholders and funding is essential. The Reach, Effectiveness, Adoption, Implementation, Maintenance (RE-AIM) Model for the evaluation of the effectiveness of interventions can be a proper framework to estimate the future impact of the incentive program. RE-AIM elements follow a sequence beginning with adoption and reach, followed by implementation and efficacy and finally maintenance. For this, it is important to establish the difference between the weight loss program and the incentive program. Please find attached a table summarising the possibilities of RE-AIM components being met by the incentives included in the weight-loss program.

A model of behavioral changes to explain the effect of the intervention
Obesity is a chronic disease that, as hypertension or diabetes, must be treated for life. The article does not mention the components of the future intervention. Assuming that the intervention will be guided by current Obesity Clinical Practice Guidelines, lifestyle and behavioral changes (dietary and physical activity) and the use of obesity medications should be included. Behavioral changes are one of the cornerstones of the obesity management.
The Stages of Change Model initially developed based on the experience of smokers who quit, propose that change in behavior occurs continuously through a process with several steps. Each step has a duration: pre-contemplation (6 mo), contemplation (6 mo), determination (1 mo), action (6 mo) and maintenance (6 mo). There are strategies that are more effective for each stage of change, and the goal is to reach the maintenance, the ideal stage of behavior. In the maintenance stage, people have sustained their behavior change (e.g. weight loss) for more than 6 months and intend to maintain the behavior and avoid relapsing. Monetary incentives for 9 months as is proposed in the article could accelerate the initial steps but do not guarantee maintenance of behavioral changes that effectively let them maintain a healthy weight. This is related to the comments about sustainability mentioned above. Please find attached a figure depicting the Stages of Change Model.

Obesity as a complex disease
The present article is exploring the possible components of program but fail to mention the complexity of factors involved in the process of implementing this strategy to increase the adherence. The obesity per se is a complex disease with multiple pathways controlling individual feeding behavior. Also, there are physiological adaptations occurring after weight loss such as changes in body composition, hormonal environment, energy expenditure, and control of food intake that predispose to regain the weight loss. The only options showing sustainable results and low food availability and access that have occurred during crisis and famine periods, and the forced modification of gastrointestinal anatomy by bariatric surgery. To offer money to participants to stimulate their adherence need to be tested in the future compared with a group without incentives in a well-designed randomized clinical trial. It is possible that the strategy can be effective for a short time but difficult to maintain.

CONCLUSION
The conclusion is very direct and only mentions that incentives seem to be a strategy widely accepted by the diabetic population of Lima. This is an excessive generalization considering the sampling and the observations I made in the last paragraph on the background.
Remembering: Participants question the usefulness of the incentives to increase the willingness to participate since 92% of them reported agreeing to participate in a weight loss program without incentives. Also, 19% considered that is not good to receive money for taking care of the own health.

LIMITATIONS
Based on my previous observations, review the limitations again. I leave it to your discretion to reconsider the need to include some additional aspect.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Author Response 22 Sep 2018
Harold Akehurst, Great Western Hospitals NHS Trust, Swindon, UK Dear Dr Nieto-Martinez, Many thanks for taking the time to offer your review of our paper. We have revised the manuscript and hope that we have addressed your concerns. We have revised the references in the introduction largely in accordance with your suggestions. We did not revise refs 5-7 (in the version you reviewed), which are intended to support the assertion that self-management programmes do incorporate lifestyle changes, rather than their efficacy, which is the subject of the preceding sentence. You express very understandable concern that anthropometric and laboratory data are not reported. We agree entirely that these are desirable. Our study did not have access to medical records, and to do so would have required significantly more manpower than could justifiably been expended for a small pilot study (the setting does not benefit from electronic medical records). The decision to restrict to a questionnaire was pragmatic. We have nonetheless elaborated on this limitation in the revised manuscript. Although the absence of independent verification of diabetes disease status does mean that theoretically respondents might not have actually been diabetic, we do consider this unlikely given their recruitment from a diabetes clinic, and that 95% were able to state that they were taking antidiabetic treatment. We are especially grateful for your scrutiny of Table 3, in which there were indeed a number of numerical errors which have been corrected. We have also corrected an error which you noticed in Table 3 about participants suggested methods for health control. You ask specifically about the response "Avoid appointments", which I must confirm is correctly rendered. The logic is the participant's own. You kindly offer three additional topics for discussion, some of which you will find included in the revised manuscript. Your suggestions about the logistics of weight loss interventions, the RE-AIM model, and the holistic management of obesity are clearly critical to the design and evaluation of an intervention such as that planned by our research group. Nevertheless, it is difficult to integrate these issues into a discussion of the data which our questionnaire study has generated. The discussion section is already as long as the introduction, methods and results combined, and out of concern not to 'bury the lede' we are reluctant to expand it beyond the scope of our exploration of participant preferences. Again, we are very grateful for your kind review.

Luzia Valentini
Institute of Evidence-based Dietetics, University of Applied Sciences Neubrandenburg, Neubrandenburg, Germany Sara Ramminger University of Applied Sciences Neubrandenburg, Neubrandenburg, Germany The present report deals with a patient survey on potential monetary incentives to increase adherence in future weight loss programs to address type 2 diabetic in poorer section of the population in Peru. To address the topic is thoughtful as well as important and it clearly deserves publication.
Nevertheless, the survey itself leads to no clear solution. On one hand it shows clear results on the missing attractiveness of the deposit-contract scheme and that cash/bank transfer is preferred over vouchers. But on the other hand it offers also important insights into the prejudices towards monetary incentives as well as it potential benefits. Interestingly, the participants themselves addressed the potential short-lived nature of monetary incentives during the weight loss without addressing incentives to maintain weight afterwards. These results, therefore, are ambiguous and not disputed in the discussion section. The discussion and conclusion parts do not sufficiently address major points of the result section. In hindsight the authors might have designed the questionnaire differently, offering more options, possibly also ones that address intrinsic motivation and not only extrinsic motivation, as shown in this report.

Major points:
Ad 2) Study design: Anthropometric data of the participants are missing. What was the mean BMI (standard deviation)? As BMI was not an inclusion criterion: How many respondents were in the normal weight, overweight, obese I, II and III range? ○ There was a bias in the selection of participants, which was addressed in the discussion section. Apparently those who intended to lose weight agreed more often to participate in the survey as seen by the 94 %age who agreed to the unincentivised weight loss program and the skewed, high percentage of participating women. The selection bias might be unavoidable but means, that the researchers addressed "the converted" in the survey and not those who do not want to lose weight but, nevertheless, might motivated by the monetary payment. It might also mean, that motivation to participate in a weight loss program might not be achieved by alone by "money for kilos". We miss the dispute on this issue in the discussion part. It there a way to reach the so far unreachable group and can it be achieved by monetary incentives at all? ○ Weight loss is not linear but more at the beginning and less later on. The main weight loss is expected to happen in the first 3 months. Therefore, the rationale for the bi-weekly payment for 1 kg weight loss is not self-explanatory. It might demotivate at the beginning (when more than 1 kg is lost in the 2-week period) and at the end of the program (when weight loss slows down). Participants with higher BMI and more kg of potential meaningful weight loss can "earn" more money than participants in the overweight range, in whom a weight loss of 5-10 kg is adequate and this should be achieved in max. 20 weeks = 5 months according to the suggesting scheme (which is in line with international guidelines). Meaning that it is achieved prior to the program end. Is this motivating? We miss the discussion about these issues in the discussion part. Results mention that 85% would participate in an incentivized weight loss program, but even more (92%) responded that they would participate in an UNincentivised weight loss program anyway (page 3, right column, last two lines). This result should be mentioned in the abstract (and discussed in the discussion part).

○
Who is going to provide the monetary incentive? Is it realistic that the money will be available in long-term? Also this should be mentioned somewhere in the manuscript.

Ad 3) Methods -details:
The main questionnaire is only available in the Spanish language -translation to the English language would be helpful

Ad 6) discussion
The discussion does not address all major results,. E.g. Results mention that 85% would participate in an incentivized weight loss program, but even more (92%) responded that they would participate in an unincentivised weight loss program (page 3, right column, last two lines). 14 participants (19%) further considered that it was not good to receive money for taking care of their own health and one explained that it was like "selling ourself" (page 7, left col, para 2, L3-109). Especially for women, being poor and "selling yourself" implies critical and serious connotations. Surprisingly, only 16 participants thought that payment was a good idea. We think these are rather unexpected and important results, which were not addressed in the discussion part. Furthermore, five participants addressed that payment will achieve only short-term results and this was also not mention in the discussion section. It is opening an important discussion if "payment per kilo" or incremental payment DURING weight loss is indeed a promising solution or if other models might be more promising (payment AFTER achieving weight loss goals or payment (or reduction /or extra money for health care costs) during the maintenance period and on longterm. We miss discussions and critical reflections on these issues in the discussion part.
Also the conclusion is irritating and does not reflect the results.

Minor comments:
Abstract: The exact formulation of objectives is missing.

○
Result on the 92% that responded that they would participate in an unincentivised weight loss program (page 3, right column, last two lines) should be added. ○ Add "each" between "incentive for" and "1 kg", otherwise it is unclear if the incentives incremental and paid every 2 weeks and not as one payment after 9 months. ○ Conclusion is irritating and does not reflect results (see also above ad 6.) ○ Introduction: First paragraph focusses mainly on sustainability of weight loss, which is not addressed by the survey. Reading the title and the manuscript as a whole, the main aim as to our understanding was to investigate the attractiveness of monetary incentives in weight loss programs, whereas defining the demographic, social, cultural and health characteristics was secondary and only accessory (and therefore rather rough). Therefore, the phrasing of the aims is irritating to us. ○ It is unclear if the incentivized weight loss program is intended for research purposes first and for integration in a regular health care program later on OR is it intended for research purposes only OR is it intended to start straight with a regular health care program? Information on this would be helpful.

Results:
As already mentioned in major comment 1) we miss the BMI/body weight data of the participants.
○ Page 4, left col, L4-5: wording "because they thought 9 months was a long time to avoid craving" is unclear ○  Table 3: it would have been also interesting to know the reason WHY the participants found it difficult to implement the health control measures, if they tried and even if they have not yet tried. We miss this point in the limitation section of the discussion. We also wondered about some answers on the methods for weight loss control. Why are "avoid appointment" and "foot care" listed? A short explanation would be helpful ○ Discussion: See also major point 6. Major issues are not discussed and the conclusion is irritating. ○ Page 7, left col, last para: What is meant by "the second method"? ○ The discussion remains vague, for example it is nice to read about the different theories about behavior change, but where is the link to the results or the conclusion?