The EHCV scheme provides financial incentives to older adults for choosing a range of primary healthcare services in the private sector. On one hand, as expected, voucher use was associated with increased uses of private primary care services (Figure 2). On the other hand, the EHCV was not significantly associated with public GOPCs, hospitalization, and A&E utilization. Our findings were consistent with local studies [22, 23]. Descriptive analysis also demonstrates that the utilization rates of public services were much higher than private ones, among both EHCV and non-EHCV users (Table 2). These findings imply that the multimorbid older adults still go for public healthcare service, despite the use of EHCV. This might be attributable to the inadequate monetary amount of voucher (HKD 2000 or USD 258.0) , leading to impractically have long-term management for chronic diseases in the private sector.
Ironically, voucher use was associated with higher public SOPC attendance. One explanation is that with more private family medicine and general practitioner visits by voucher users, more conditions requiring specialist care may have been diagnosed or more referrals for specialist care were made due to health screening. The multimorbid older adults may prefer heavily subsidized SOPCs over self-financed private specialist services (Table 1) . Therefore, the public sector remains the major healthcare providers for older adults with multiple chronic conditions.
In Hong Kong, pneumococcal vaccine is available at a relatively low cost in the private sector due to governmental subsidy, whereas the public sector has not provided free-of-charge pneumococcal vaccine to older adults until 2017/18 . Older adults were likely to receive the vaccine from private healthcare providers and pay for it with the EHCV. This may explain the reason why voucher use was positively associated with pneumococcal vaccination (Figure 2).
Nonetheless, the use of the EHCV was not significantly associated with seasonal influenza vaccination (Figure 2). Unlike the pneumococcal vaccine, the influenza vaccine has been freely available for older adults in the public sector at the time of the survey. Hence, older adults might have received the influenza vaccine from public healthcare providers, in order to save the EHCV which is of limited amount for other purposes.
With the financial incentive provided by the EHCV, older adults should have higher purchasing power and thus theoretically more choices of private healthcare service. The quality of private services should be improved through market competition . However, instead of choosing a higher quality of care, we found that patients may end up doctor shopping in the private sector. This may be related to the moral hazard of overutilization, perverse incentives, information asymmetry between patients and healthcare providers, or a lack of a well-established family doctor system in Hong Kong . As shown by our results, EHCV users were less likely to visit the same primary care providers for chronic disease management (Figure 2). Without a common electronic health record sharing system, the EHCV scheme may have aggravated the fragmentation of clinical information and management approach in the private sector, rather than promoting continuity of care and the concept of family doctor.
An evaluation review suggests a re-design of the EHCV scheme to promote preventive services, chronic disease management, and continuity of care . Specifically, the voucher should be designated for (i) preventive care for early detection and treatment and for (ii) chronic disease management, instead of a broad service coverage under the current scheme. The findings of this study support the suggestions.
Limitations and Further Study
Since this study only used secondary data, methodological limitations were related to study design and data availability from the Elderly Health Care Quality Survey. First, we recruited the participants from GOPCs, SOPCs, and geriatric day hospitals in the public sector. We did not include those who have shifted from public to private sector completely due to EHCV incentive. Second, clinical information including disease severity and reasons of consultation was not available in the data source and therefore, we were unable to adjust for any unmeasured confounders or differences in disease severity between voucher users and non-users which may have affected our findings. Third, both predictor and outcomes were binary variables. Future studies should adopt a more detailed record of measurements (interval and ratio), such as the various amount of voucher use and frequency of service use, for more rigorous data analysis in the future. Further studies should evaluate the effects and cost-effectiveness of the EHCV Scheme.