JMIR Publications

ABSTRACT

Background:

The increasing use of willingness to pay (WTP) to value the benefits of eHealth offers an opportunity to estimate the average WTP for eHealth, and to study the heterogeneity between the valuation of different modalities of eHealth among different populations.

Objective:

We aimed to systematically review WTP studies about electronic health (eHealth), obtain an overall estimation of WTP for eHealth, and examine related factors.

Methods:

We searched Medline, CINAHL Plus, PsycINFO and EconLit databases from their inception until June 22, 2020 for WTP studies about eHealth. Random-effects meta-analyses were performed to calculate the mean WTP for eHealth (in US$2019). Meta-regression analysis was undertaken to examine factors affecting WTP for eHealth.

Results:

We reviewed 27 relevant research articles (representing 32 studies) and obtained estimations of the mean WTP for eHealth (one-time payment: 16.80US$; monthly payment: 10.06US$/month). Comparisons between included studies showed that the WTP value was affected by the percentage of female respondents (β=1.31, p<0.0001), percentage of respondents that completed college education (β=1.52, p<0.0001), GDP per capita ('000US$) of the country where the study was conducted (β=0.05, p<0.0001), modality of eHealth, and the methods used to elicit WTP. For the modality of eHealth, comparing to website, people showed lower WTP for asynchronous communication (β=-2.69, p<0.0001), and higher WTP for medical devices (β=0.32, p=0.0008) and synchronous communication (β=0.38, p<0.0001). As for the methods used to elicit WTP, comparing to open-ended format, double-bound dichotomous choice format (β=2.57, p<0.0001), payment scale format (β=1.60, p<0.0001), and those that did not report the format (β=2.44, p<0.0001) were shown to obtain higher WTP values; while the use of single-bound dichotomous choice format was associated with lower WTP values (β=-0.25, p=0.03). Comparing to ex ante evaluation, ex post evaluation was shown to obtain lower WTP values (β=-0.18, p=0.0002). Studies that excluded protest zero responses or excluded all zero responses reported higher WTP values than studies that included all zero responses in the analysis (β=0.37, p<0.0001).

Conclusions:

WTP for eHealth across different studies varied greatly depending on the study population, eHealth modality, and the methods used to elicit WTP values. WTP for eHealth was lower among certain populations (i.e., individuals who were male, had lower education level, or had lower income), suggesting that these populations may be disadvantaged at accessing and benefiting from eHealth, and calls for efforts to resolve the disparities. We also identified the eHealth modalities and features that were highly valued and offered suggestions for the design of eHealth interventions. From the methodological perspective, different approaches used to elicit WTP values led to significantly different WTP estimations, but no consensus has been reached about the optimal approach. Further methodological comparisons between different approaches and refinement of the approaches to elicit WTP values were warranted.