Effectiveness of Financial Incentives for a Pedometer-Based Walking Promotion Program

Physical inactivity contributes significantly to poor health and the onset of disease. Physical inactivity is also associated with severe economic burdens. Japans Ministry of Health, Labor and Welfare (MHLW) cites the provision of various health promotion incentives aimed at health-indifferent groups, which are made up of individuals less interested in health promotion. This study investigated the relationship between medical costs and a pedometer-based walking program providing monetary incentives based on daily step counts. The study sample included 16,816 citizens aged 40-75 years who lived in Takaishi City and were enrolled in the NIH from October 2016 to March 2018. The results of the ordinal logistic regression analysis showed that participation in a walking promotion program with an HPFI was correlated with a reduction in healthcare costs over 1 year in a sample of Takaishi City residents. In fact, there was a difference of 67,077 yen in the average medical cost per person per year between the walking and control groups. With 1,923 walking group participants, the total medical cost reduction was predicted to be at least 12,898,904 yen. In this study, we evaluated the effectiveness of a walking promotion program with incentives. With program context differing widely from region to region and country to country, future investigations are needed to inform the selection of appropriate incentive schemes for programs offered in other regions and countries.

4 66 homes, the city's website, and posters around the city. Located in an industrial area of Osaka, the 67 city has a population of 56,000, of which 27% is older than 65 years. Takaishi city residents who 68 wished to participate in the walking program were required to register at the city office, where 69 they received a pedometer (Activity Meter with FeliCa AM-150; Tanita Corp., Tokyo, Japan).
70 Participants could also use the pedometer application on their smartphones.

71
At their convenience, participants registered their daily step data by bringing their 72 pedometers to special machines installed at community centers, post offices, and stores 73 throughout the city. Depending on the number of steps taken each day, up to 6,000 points were 74 awarded over the course of a year (one point equaling JPY1), which could be exchanged for gift 75 certificates. As there is no private medical insurance in Japan, every resident is required to enroll 76 in one of two public health insurance systems. [6,7] This study used a claims data set from one of 77 these systems, the National Health Insurance (NIH), which covers self-employed people (e.g., 78 farmers and fishermen), retirees and their dependents, and the remaining 25% of the population.
79 This data set included the following information: ID number, the month in which domestic 80 diagnosis codes were assigned in each hospital or clinic, whether claims were inpatient or 81 outpatient, whether the claims were medical or DPC claims (medical claims form), the number 82 of days per month medical treatments were provided, domestic diagnosis codes (mapped to the 83 10th Revision of the International Statistical Classification of Diseases and Related Health 84 Problems), and the medical fees scale (medical expenses). These data were selected in 85 compliance with rules governing the use of personal information and then prepared for analyses.
86 Each individual was given a temporary ID (standardized for medical cost and pedometer data) 87 that was used to link same individuals during analyses. Individuals with registered step count 88 data were considered program participants, whereas those without such data were not considered . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint 89 participants. The daily step data were used for analyses, and the first month with registered step 90 data was considered the first month of program participation. This research protocol was 91 approved by the Ethics Committee of the Osaka City University School of Medicine.

93
This study was designed to assess the impacts that the program participation had on 94 healthcare costs. When setting up the control and walking groups, attention needed to be paid to 95 factors other than program participation that could affect medical costs in both groups. In this 96 case, there was no data on factors that might be expected to affect the outcome (i.e., medical 97 history, smoking history, body mass index, etc.). Therefore, to adjust for the background factors 98 affecting healthcare costs of the walking and control groups, we adopted a method to reduce bias  Table 2. There were no significant demographic 121 differences between the walking and control groups． . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint 122 123 124 Figure 1 shows the program participants' step registration rate. Results showed that 86.4% of the 125 participants registered their steps for more than 80% of the days in the registration period. Figure   126 2 and Table 3 shows the results of the ordinal logistic analysis, which indicated that participating 127 in a program with a health-promoting financial incentive (HPFI) was associated with a reduction 128 in healthcare costs over a 12-month period.

129
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint 9 151 controlled trial to evaluate the effects of an HPFI that provided monetary incentives according to 152 the number of aerobic exercise minutes recorded by a weekly pedometer over a 4-week period. 153 The results showed that the incentive group had a higher level (1.7 times) of weekly exercise 154 than the control group.
[9] Therefore, the researchers concluded that linking a moderate monetary 155 incentive to aerobic exercise minutes is an effective and potentially cost-effective approach to 156 increasing physical activity.
[9] Although that report had the advantage of using a randomized 157 controlled design, it had a small sample size of only 51 subjects. In contrast, this study was 173 is concern that an HPFI's effect decreases rapidly after the incentive is withdrawn [15] and that . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint 175 suggest that although health promotion interventions are associated with short-term behavioral 176 changes, the long-term effects of such interventions require further investigation.

177
The current walking promotion program with an HPFI was conducted for 2 years and 6 178 months, from April 2016 to March 2019. Although citizens were given incentives based on their 179 daily step count, there was no penalty for not registering steps. Based on this design, it could be 180 predicted that participants inevitably would register their steps less frequently as their interest in 181 the program waned. As the total participation period depended on the registration start date, this 182 study's analyses were limited to only 12 months from the registration start date. Of the 183 participant sample, 86.4% registered their daily step count on more than 80% of the days in a 184 year, which is considered a very high participation rate.

185
This study has some limitations. As the participants voluntarily chose to participate in the 186 intervention, the results may not apply to the general population. Moreover, as the control group 187 did not register a daily step count, it was impossible to determine whether any changes in their 188 step count or medical expenses occurred during the study period.

190
These results suggest that a walking promotion program with an HPFI could help reduce 191 healthcare costs. For incentive measures to work effectively, the rewards must be attractive to 192 participants. If the rewards are too small, the incentive will not work. Conversely, the larger the 193 rewards, the higher the program expenses will be as the rewards are paid to participants. This 194 program's incentives were based on the HPFI guidelines from the MHLW as well as incentives 195 granted by similar policies implemented in Japan. With program context differing widely from 196 region to region and country to country, future investigations are needed to inform the selection . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 8, 2022. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint activity in a UK behavioral incentives program. npj Digit. Med. 2019;2: 91.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 8, 2022. ; https://doi.org/10.1101/2022.02.07.22270574 doi: medRxiv preprint