Background: Screening for diabetic retinopathy (DR) has been proven effective and cost-effective in preventing blindness. Hong Kong (HK) has a mixed health care economy and, before this study, there was no systematic screening for DR. The optimal screening interval for DR screening is controversial with some countries extending it to 2 or more years. Risk algorithms tailor screening intervals to an individual but the safety of a 2-year interval and the validity of the Iceland Risk Algorithm (IRA) are uncertain in HK. This study assesses the impact of charging a co-payment for DR screening in the public sector, evaluates its cost-effectiveness and examines the use of an algorithm to determine optimal screening intervals for subjects in HK.
Methods: A randomized controlled trial (RCT) was conducted with subjects with diabetes from two general outpatient public clinics randomized to free screening or pay screening with a co-payment of HK$60. Cost-effectiveness analysis used a Markov cohort model to compare these with opportunistic screening. Incremental cost-effectiveness ratios (ICERs) were calculated and one way and probabilistic sensitivity performed. Subjects were followed up for two years to examine the safety of a 2-year screening interval and to test the validity of the IRA in predicting sight threatening diabetic retinopathy (STDR). A new prediction model using cohort data was developed using logistic regression and tested in a similar fashion.
Results: After randomization, 1316 in the free and 1277 in the pay group agreed to participate. Uptake of screening was 88.5% (1165/1316) and 82.4% (1052/1277) in free and pay groups respectively (Pearson chi=19.74, P<0.001). Being in the pay group was associated with lower uptake of screening (OR=0.59, 0.47 to 0.74) and lower detection rates of DR (OR=0.73, 0.60 to 0.90) after adjustment.
From the societal perspective, pay systematic DR screening rather than opportunistic screening gives an ICER of HK$94,630/QALY gained. Free rather than pay systematic screening, had an ICER of HK$199,741/QALY gained. Probabilistic sensitivity analysis showed when willingness to pay for a QALY was HK$186,186 or more, free systematic screening had the highest probability of being cost-effective.
The 2-year cumulative incidence of STDR was low for those with no DR (2.9%) and those who developed STDR did not experience severe visual loss during follow up. The IRA had good discrimination for identifying STDR, but significantly lacked calibration. The new prediction model improved discrimination and calibration compared with the IRA.
Conclusion: A number of people in the pay group did not uptake screening and appear to be higher risk cases. The inverse care law appears to operate even with this relatively small co-payment. From the societal perspective, free systematic screening was more cost-effective than pay within the WHO threshold of 1 x annual per capita GDP (HK$338,520) for a QALY. Free systematic screening can be considered the most cost-effective screening strategy from the societal perspective. We could tailor screening intervals according to individual risks using a new prediction model which appears safe and efficient but which requires further testing with follow-up data. / published_or_final_version / Public Health / Doctoral / Doctor of Philosophy
| Identifer | oai:union.ndltd.org:HKU/oai:hub.hku.hk:10722/196489 |
| Date | January 2013 |
| Creators | Lian, Jinxiao, 連金晓 |
| Contributors | Wong, DSH, Johnston, JM, McGhee, S |
| Publisher | The University of Hong Kong (Pokfulam, Hong Kong) |
| Source Sets | Hong Kong University Theses |
| Language | English |
| Detected Language | English |
| Type | PG_Thesis |
| Rights | The author retains all proprietary rights, (such as patent rights) and the right to use in future works., Creative Commons: Attribution 3.0 Hong Kong License |
| Relation | HKU Theses Online (HKUTO) |
Page generated in 0.0055 seconds