Mobile phone use and willingness to pay for SMS for diabetes in Bangladesh

Abstract

Background

Mobile phone SMS is increasingly used as a means of communication between patients and their healthcare providers in many countries of the world. We investigated mobile phone use and factors associated with willingness-to-pay (WTP) for diabetes SMS among patients with type 2 diabetes in Bangladesh.

Methods

As part of a randomized controlled study, in 515 patients with type 2 diabetes, socioeconomic status, mobile phone use, WTP for diabetes SMS, anthropometry and HbA1c were measured. Multivariate regression was used to identify factors associated with WTP.

Results

The median (interquartile range [IQR]) of WTP for diabetes SMS was 20 (45) Bangladesh Taka (BDT) (1 BDT = 0.013 US$). WTP was significantly higher for males [OR 2.4, 95% CI (1.0–5.7)], patients with household income >50 000 BDT [4.6 (1.1–20.4)] and those with primary education [5.6 (1.2–26.6)] and secondary and higher education [5.2 (1.4–19.6)].

Conclusions

The high proportion of mobile phone use and WTP for diabetes SMS are encouraging as possible strategy to use such technologies and deserve further evaluation.

Introduction

The diabetes epidemic is progressing rapidly with disproportionately higher rates in developing countries. The number of individuals with diabetes is expected to double by 2030 reaching more than half a billion globally and imposing huge financial burdens.¹ Bangladesh is one of the hardest hit countries, and its rather fragmented healthcare system has proved to be unprepared to deal with this emerging crisis. Innovative approaches using information technology and mobile health (mHealth) might be an option where traditional approaches have failed to deliver sustainable health attention to individuals with diabetes.

Mobile phone SMS is increasingly used as a means of communication between patients and their healthcare providers in many countries of the world. Several studies have shown that SMS counseling/follow-up can improve patient's behavior and health outcome.^2–5 However, the potential use of SMS in clinical settings in developing countries has not been well established. The feasibility of implementing such technologies lack strong evidence, as well as cost-effectiveness and sustainable payment mechanism or business model that can be scaled-up.^6–12 Information about how patients value or perceive the use of mobile phone SMS to improve diseases management may be a valuable aid in implementing such methods in primary care.

Willingness to pay (WTP) is a widely used approach to valuing services not available to consumers by constructing a hypothetical but realistic scenario and measure their maximum WTP if the services existed.^13–15 Patients in Bangladesh traditionally have to pay fees for physician's visits and healthcare services in both government and private facilities. It is therefore likely that patients would be willing to pay for a new service that they consider beneficial. Patients with chronic conditions like type 2 diabetes require lifelong follow-up, and the use of SMS could be an option to reduce physician's visits, to provide appointments and medication reminders and to facilitate low-cost communication between primary care physicians and patients. There is a lack of information about the costs, affordability, patient's WTP, benefits to the users and cost-effectiveness of the services. This paper aims to assess mobile phone use and associated costs for patients with type 2 diabetes in Dhaka city and their WTP for a mobile phone SMS service for diabetes care.

Methods

We conducted a WTP study as part of a randomized controlled trial on mobile phone intervention for increasing adherence to treatment for diabetes in Bangladesh. Detailed methodology of the study design has been published elsewhere.¹⁶ In brief, 515 patients with type 2 diabetes attending the outpatients department (OPD) of Bangladesh Institute of Health Science (BIHS) hospital in Mirpur, Dhaka, were recruited from September to December 2013. The inclusion criteria were as follows: both male and females, aged 18 or above, diagnosed as type 2 diabetes by BIHS physician according to WHO criteria, on exclusive oral medication therapy, living in Dhaka city, owning a mobile phone and willing to provide written informed consent. Data were collected through face-to-face interviews using a structured questionnaire at the OPD of BIHS hospital by a team of three trained research assistants, a research officer and a research physician. All consecutive patients meeting the selection criteria were included in the study. The study was approved by the ethics committee of the International Center for Diarrhoeal Diseases Research, Bangladesh (icddr,b) and received ethics committee waiver from Ludwig-Maximilians University (LMU) and Bangladesh Institute of Health Science (BIHS).

Variables and measurements

The questionnaires contained questions on socioeconomic characteristics (age, sex, marital status, education and income), mobile phone use, SMS use, WTP for SMS, health seeking behavior (primary point of care, distance to health center, travel costs and number of medication use), self-reported complications and result of blood glucose tests. Participants were asked through an open-ended question how much they would be willing to pay per month to receive SMS related to diabetes care, if it was available.

WTP was defined as the monthly amount of money each participant would be willing to pay if an SMS service for diabetes was available, in addition to regular medical care. Participants were informed that the SMS would be related to diabetes care. None of the participants had any experience of such services and thus were uncertain about the usefulness of such service. We used contingent valuation methodology (CVM) that is widely used to quantitatively measure patient WTP utilizing open-ended questionnaire, as it is more flexible in terms of amount of money WTP for SMS and provides a high degree of individual impreciseness than bidding game, payment card or multiple choice questions to avoid a starting point bias.¹⁴

Data analysis

The mean ± standard deviation (SD) and median with interquartile range (IQR) for amount of money the participants were willing to pay were established, and relationships between WTP and participant's characteristics were assessed. WTP was transformed into two categories: Zero WTP versus WTP. We conducted χ² tests to assess the significance of WTP responses and associated factors and Mann Whitney U tests and t-tests to show the difference between continuous variables such as WTP groups. Univariate logistic regression analysis was used to assess the association between WTP and individual variable. Factors with statistically significant association with WTP in univariate analysis were included in the final logistic model controlling for other variables. SPSS version 20 was used for data analysis. A P-value of <0.05 was considered statistically significant.

Results

The participants’ mean ± SD age was 50.0 ± 10.1 years; 55.9% were females. The monthly household income was 39 200 ± 36 000 Bangladesh Taka (BDT) or 504.99 ± 463.77 $US (1 BDT = 0.013 $US); 40.6% reported complete college or above education level and 68% had a family history of diabetes.

Mobile phone use and WTP

All participants in this study owned a mobile phone. About half of the participants reported to be able to read/retrieve SMS, while only 36.1% could send SMS. The median (IQR) years of mobile phone use among participants was 7 (6) years, which was significantly higher among males. A greater number of participants reported to read or send one SMS monthly (46.6%) and very occasionally (36.5%) than those who read or sent SMS more frequently. Half of the participants never read any SMS, one-fifth occasionally read and about one-third reported to read all SMS or only from known people. The use of SMS was significantly higher among males (P = <0.001). The median (IQR) amount spent in a month for mobile bills was 300 (600) BDT. Males also reported spending significantly more on mobile phone bills compared with females. All participants in the study, except two females, reported their interest in receiving SMS for diabetes (Table 1).

Of the 515 participants, 268 (52.0%) expressed a positive WTP, 84 (16.3%) expressed a Zero WTP and 163 (31.6%) did not know how much to express or did not answer (missing). This left us with 352 (68.3%) for analysis of WTP. The median (IQR) WTP for SMS for diabetes in a month was 20 (45) BDT. The median WTP was around 36% of participants’ monthly physician fees, 10% of mobile bill, 8% diabetic food costs and 3% of total monthly medication costs.

Factors associated with WTP

The mean ± SD age of WTP, Zero WTP and those with no response was almost similar. Among those willing to pay for diabetes-related SMS, WTP was significantly higher among males (60.8%), participants with monthly income category 10 001–30 000 BDT (51.9%), completed secondary or higher education (57.6%), participants who had at least one visit in last 3 months (57.6%), those who paid physician fees <500 BDT (67.2%), consulted MBBS (graduate medical) doctor (61.2%) and treated at a local pharmacy (71.7%) (Table 2).

In the multivariate analysis, controlling for all other factors, WTP was significantly higher among males [OR 2.4, 95% CI (1.0–5.7)], those with household income >50 000 BDT [4.6 (1.1–20.4)] and among those with primary education [5.6 (1.2–26.6)], secondary and higher education [5.2 (1.4–19.6)]. Adding the ability to read or write SMS in the model did not change the results much (Table 3).

Discussion

Main findings of this study

This study, to the best of our knowledge, represents the first attempt to investigate the WTP for SMS-based service for any chronic, non-communicable diseases in Bangladesh. The Government of Bangladesh in line with WHO recommendations stressed the use of information technology including mHealth. In this study, except for two participants, all expressed willingness to receive diabetes-related SMS and the median (IQR) WTP was 20 (45) BDT per month, which seems reasonable for such services compared with the current expenditure for physician fees.

What is already known on this topic

mHealth is a new and developing field with numerous pilot implementations, all in an attempt to achieve the increase access to healthcare and technology. One shortfall to date with this approach has been the sustainability of such projects when the funding withdraws—resulting in recognition of the need for financial self-sufficiency if the industry is to render long-term benefits. A study on patient perceptions of a mHealth service finds that the sustainability of a service is dependent on user willingness to incur the usage costs but excludes further WTP investigation.¹⁷ The approach to their research was to use the Technology Adoption Model that focuses on usage of technology rather than payment for the technology. A mHealth surveys in six countries reported on interest and WTP for four hypothetical products and provides a high-level view of perceptions of the general population, but without tailoring a specific product to a specific condition, the risk of ‘hypothetical bias’ is high.¹⁸

WTP is a measure of analysis used in the field of health economics to understand the willingness of health users to pay for specific treatments or services and establish how new technologies, if available, would be valued by the consumers.^19–23 In this study, WTP is specific to the individual and should be considered in an economic context—it does not mean that the user will always be ‘happy’ to pay for the service, but they would rather pay for the service at a certain price rather than not have the service.²⁴ When assessing this ‘trade-off’ between the money paid and the mitigation of a health risk which is always hypothetical, it is important to consider the limitations and difficulties when using the WTP method.²⁵ Customers having full information about the service that they are being asked to evaluate, as well as assuming that the customer is fully rational, is essential. The accuracy of the findings may be limited by the respondents not having a full understanding of the changes to the proposed service.²⁵

While the hypothetical nature may introduce greater variance, this does not in itself introduce a bias. The concern that WTP is a function of the economic and social status of the respondent is also intrinsically part of the measure. While some of these concerns can be mitigated through careful survey and sampling design and implementation of the WTP best practices that have been developed for use in multiple fields, more pressing are the biases that have been identified in WTP studies to date.²⁶ Different sources of bias may affect the results of the interviews, and these have been termed ‘response effects’.²³ Response effects may be due to respondents answering through motivations other than simply providing a realistic view of their WTP, such as trying to conform to the expectations of the interviewer. Despite a recognition of various shortfalls of the WTP method, the ‘approach is now firmly established in the research community’. Due to the widespread use, WTP has been shown to display ‘feasibility of use in a broad range of different technologies and diseases’.²³ WTP is principally a quantitative strategy and to better contextualize the results qualitative interviews might be helpful.

What this study adds

In our study, those in the age groups of 40–49 and 50–59 years showed higher WTP than other age groups and WTP decreased slightly with age. Older respondents are less well-off than their younger counterparts, and this could be the reason for less WTP. In the multivariate analysis, only those with household income in the range of 20 001–30 000, 30 001–40 000 and above 50 000 were significantly associated with WTP. An increase in education correlates with an increase in the proportion of income earners per household and had an impact on WTP, contrary to the literature's assertions in this regard.²⁷ Those with relative higher income had more WTP that is aligned with the expectation.^23,28,29 Despite being willing to pay a lower absolute value for SMS, it is interesting to note that the WTP was a large proportion relative to their travel costs for health care (93%), monthly physician fees (36%) and monthly mobile bills (10%). This may be due to being less satisfied with the current level of health care that they are receiving or willing to use SMS as an alternative mechanism for communication with their physician.

Almost one-fourth of our participants (23.8%) were not willing to pay for SMS services. In a previous study in Norway measuring patient's WTP for electronic communication with the general practitioners found 48% not willing to pay. Another study by Anand found 63% of participants not willing to pay for email communication with their physicians which is much higher than our findings. As our participants did not have experience of the SMS service for diabetes which is considered as a new technology, it is likely that their impression towards such services will be positive or neutral in the absence of negative information before the trial in a hypothetical scenario.

Limitations of this study

One potential limitation of our study was that it was conducted among newly diagnosed type 2 diabetes patients on oral medication therapy and willing to participate in a clinical trial. This group might be more willing to invest in their personal health than the general population. Also, the study participants in this trial all had access to a mobile phone which might be different from the diabetes population in Bangladesh. There is a large divide in terms of healthcare service at public and private facilities, and therefore, the findings of the study may be more relevant in similar not-for-profit private sectors. Another limitation of our findings is that we did not correct for the multiplicity effects in the modeling. While the research assistants informed the participants about the SMS service and possible benefits to the user, the respondents in our study had no experience of similar services. The perceptions of a user based on an actual implementation could therefore differ. This research is intended to assess the potential for generating revenue from an SMS service and developing a sustainable business model for mHealth in an emergent economy. While WTP does result in a financial figure, this should not be considered as a final sales price as it is specific to the proposed service used for the study.

Conclusion

Results of this study demonstrate that a great majority of individuals with type 2 diabetes in an urban area of Bangladesh are willing to receive SMS for diabetes and to pay a small amount for such a service. Considering the huge number of diabetes patients and the low mobile phone rates in Bangladesh, a self-sustained business model for basic mHealth services for chronic diseases is therefore feasible in this and potentially other low-income countries.

Funding

This research study was supported by the Center for International Health (CIH), Ludwig-Maximilians-Universita¨t (LMU), Munich, Germany and icddr,b.

Authors’ contributions

S.M.S.I. was responsible for study design, implementation, interpretation of data and wrote the manuscript. A.L. contributed to study design, interpretation of data and reviewed and edited the manuscript. U.F. and J.S. contributed to study design and reviewed the manuscript. L.N. and R.H. contributed to interpretation of the data and reviewed and edited the manuscript. S.M.S.I. is the guarantor of this work, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis.

Prior presentation

Parts of this work were presented in abstract form at the International Symposium on Non-Communicable Diseases in Developing Countries, Munich, Germany, 22nd March 2014.³⁰

Acknowledgements

icddr,b acknowledges with gratitude the commitment of DAAD, BMZ and Exceed to its research efforts. The authors thank Professor Fabio Zicker, Senior Visiting Professor, International Health—Center for Technological Development in Health (CDTS) at Fiocruz Foundation, Rio de Janeiro, Brazil for editorial assistance and Pear Hossain, Statistical Officer, CCCD, icddr,b for assistance in data analysis.

References