Expanding health insurance coverage in vulnerable groups: a ...

Health Policy and Planning Advance Access published September 2, 2010 Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine ß The Author 2010; all rights reserved.

Health Policy and Planning 2010;1–12 doi:10.1093/heapol/czq038

REVIEW

Expanding health insurance coverage in vulnerable groups: a systematic review of options Qingyue Meng,1* Beibei Yuan,1 Liying Jia,1 Jian Wang,1 Baorong Yu,1 Jun Gao2 and Paul Garner3

*Corresponding author. Center for Health Management and Policy, Shandong University, Jinan, Shandong 250012, China. Tel: þ86 531 8838 2692/8838 2693. E-mail: [email protected]

Accepted

30 March 2010 Vulnerable groups are often not covered by health insurance schemes. Strategies to extend coverage in these groups will help to address inequity. We used the existing literature to summarize the options for expanding health insurance coverage, describe which countries have tried these strategies, and identify and describe evaluation studies. We included any report of a policy or strategy to expand health insurance coverage and any evaluation and economic modelling studies. Vulnerable populations were defined as children, the elderly, women, low-income individuals, rural population, racial or ethnic minorities, immigrants, and those with disability or chronic diseases. Forty-five databases were searched for relevant documents. The authors applied inclusion criteria, and extracted data using pre-coded forms, on contents of health insurance schemes or programmes, and used the framework approach to establish categories. Of the 21 528 articles screened, 86 documents were finally included. Descriptions about the USA dominated (72), with only five from Africa, six from Asia and two from South America. We identified six main categories: (1) changing eligibility criteria of health insurance; (2) increasing public awareness; (3) making the premium more affordable; (4) innovative enrolment strategies; (5) improving health care delivery; and (6) improving management and organization of the insurance schemes. All six categories were found in the literature about schemes in the USA, and schemes often included components from each category. Strategies in developing countries were much more limited in their scope. Evaluation studies numbered 25, of which the majority were of time series design. All studies found that the expansion strategies were effective, as assessed by the author(s). In countries expanding coverage, the categories identified from the literature can help policy makers consider their options, implement strategies where it is common sense to do so and establish appropriate implementation monitoring.

Keywords

Health insurance, coverage, expansion, vulnerable groups

1

Downloaded from heapol.oxfordjournals.org at World Health Organization on February 23, 2011

1 Centre for Health Management and Policy, Shandong University, Jinan, China, 2Centre for Health Statistics and Information, Ministry of Health, Beijing, China and 3Liverpool School of Tropical Medicine, Liverpool, UK

2

HEALTH POLICY AND PLANNING

KEY MESSAGES

Health insurance coverage is limited in low- and middle-income countries (LMICs) and strategies for expanding coverage are needed.

A wide range of strategies for expanding health insurance coverage are available, especially in the USA for vulnerable groups.

LMICs can learn from the existing experience in how the coverage of health insurance can be expanded.

Background

(1) What are the options for expanding health insurance coverage to the vulnerable population derived from the global literature and where have they been tried? (2) What evaluation studies have been done and what did they show? We then reviewed the literature using standard methods to address these questions and this is what we report here. We judged that this information could: (a) help policy makers consider all options, and allow a discussion of the relevance and feasibility of these strategies for implementation in particular contexts; and (b) provide policy makers with an

Methods Criteria for inclusion of studies Location This review was confined to regions where health insurance programmes were established, and where the government or other agencies were modifying programmes to expand coverage of vulnerable groups. Policies Any report of a policy or strategy to expand insurance coverage was included, including overviews of policies across several countries. We excluded opinions, discussion papers and policy documents that did not refer to actual implementation of expansion strategies in specific countries. Target groups Vulnerable groups were defined as children, the elderly, women, low-income individuals, rural populations, racial or ethnic minorities, immigrants, and population with disability or chronic diseases. Study designs We included reviews, policy descriptive articles, observational studies, experimental studies or economic modelling studies. We excluded opinion pieces, letters, news, commentaries, editorials, bibliographies, meeting abstracts and background literature written in the introduction to specific projects. Language We included articles written in English or Chinese. Articles identified in English databases written in another language but with an English abstract were screened and included if they met the inclusion criteria.


Governments providing basic health care to the whole population need finance to do this. The World Health Organization recommends prepayment financing mechanisms including social health insurance, to protect against financial risk and to improve access to health care (WHO 2005). In low- and middle-income countries (LMICs), health insurance schemes often cover only a small proportion of the population (Carrin and James 2005; Hsiao and Shaw 2007). For example, in India, social health insurance covers formal sector employees, accounting for only 3% of the total population (Devadasan et al. 2006). Though community-based health insurance schemes are widely promoted in many countries in Africa, Asia and Latin America, most of them have limited coverage (Arhin-Tenkorang 2001; De Allegri et al. 2006b). Yet the people least likely to be covered are those with high health care requirements and needing financial protection (Gilson et al. 2000; Habtom and Ruys 2007; Hsiao and Shaw 2007). As there is considerable debate about how best to expand coverage to low-income groups, we considered how best this can be informed by reliable research and summaries of research. We first sought evidence from existing systematic reviews. A review by Lagarde (2006) identified one controlled before-and-after study from Rwanda of community health insurance, and no evaluations of social health insurance. A second review of community health insurance with broader inclusion criteria suggested that these schemes to date had small effects on a limited section of the population but what might work where was not apparent (Ekman 2004). As there seemed such a dearth of reliable evidence about whether schemes worked from systematic reviews, we took a step back from the ‘benefits and harms’ paradigm. We considered what questions policy makers, tasked with expanding health insurance coverage, might have that could be informed by the existing literature and evaluation studies. We identified two simple questions.

option to move forward with strategies where it was common sense to implement them, and then monitor the implementation; or, where there was sufficient uncertainty as to whether they would be effective, to implement through a more formal experimental or quasi-experimental approach, such as randomized trials, controlled before-and-after studies or interrupted time-series analysis. We also debated about whether to use studies from high-income countries, in particular the USA, where the context is so different. However, we considered that understanding policies tried in the USA may be helpful for LMICs to generate ideas and avoid similar mistakes being made, and made the review global.

HEALTH INSURANCE COVERAGE EXPANSION

3

Screening In an initial pilot programme, we randomly selected 5% of searched results, which were independently assessed by two of the authors and then each discussed to ensure the same approach was being used. Subsequently each paper identified was screened by title and abstract, and then all the literature included through titles and abstracts was searched. The retrieved full texts were screened and the final included items were identified. The process of screening was done and recorded in the EPPI Reviewer (EPPI-Centre 2008).

21528 334 excluded: opinions, comment, editorial

21194 20 934 excluded: no strategies for expanding health insurance coverage

Data extraction 260 46 excluded: strategies were not implemented 214 49 excluded: strategies were not for vulnerable populations 165

Data analysis The data analysis was guided by our review protocol. We adapted the framework approach used in the analysis of qualitative literature. As we read all articles, we grouped strategies into their corresponding categories from our initial framework. Strategies that were not grouped into the originally defined categories were described in detail and grouped into a temporary choice ‘Others’. Once the coding was completed, all reviewers and subject experts discussed the contents of all the strategies and adjusted the categories based on the exact strategies found. All strategies were finally grouped into six main categories. Within each category, in an iterative process, we developed a more detailed framework, informed by the literature we had identified, comparable to ‘sub-themes’ in qualitative analysis. The literature fell into either descriptive articles of the policy or strategy (for objective 1), or articles that included some measure of implementation success such as change in coverage (objectives 1 and 2). For the latter we also briefly summarized the evaluation design.

86 Figure 1 Application of inclusion criteria

source or location. Of the 79, most described US programmes (67, 85%), with only 11 documents (13.7%) from developing countries. Of the 86 included documents, 61 studies were descriptive articles and 25 were evaluation studies. Descriptive articles from the USA dominated (72), with a smaller number from LMICs: Africa (5), Asia (6) and South America (2). Target populations of strategies from the included documents were children, women, elderly, low-income populations, rural population, and individuals with disability or chronic diseases.

Results Selection of the articles Figure 1 shows the study selection process, with over 21 528 articles screened. Finally, 86 documents met the inclusion criteria and were included in this review (Figure 1). During this process, we obtained the full text of 654 articles. However, there were 79 marked for retrieval that, despite our best efforts, were not available: 36 were published articles in journals that we could not access in the UK or in China by any method, and the remainder were unpublished without clear reference to the

1. What are the strategy options and where have they been tried? We identified six main strategies, from our iterative process described in the methods above, each with subcategories. Table 1 summarizes the strategies and distribution of reviewed documents. Note that some schemes may have more than one document evaluating them. Below we provide an explanation of each, with some examples from the literature.


Data extraction was done by five authors. Initially, we took 5% of the results from the screening process. All these were double extracted making use of the coding form designed in the protocol. The differences in data extraction and problems in the coding form were identified and discussed by all reviewers and subject experts. Then the coding form was adjusted and improved. We used EPPI Reviewer for screening and data extraction. Relevant information was extracted from included documents. We describe the characteristics of study methods, which include data sources, time period of data, sample size and analytical methods, but not methodology quality.

4


Table 1 Main and detailed strategies for expanding coverage of health insurance schemes Strategies

Documents

1. Modifying the eligibility criteria

43

1a. Increasing income threshold for entering health insurance 1b. Expanding the categories of eligible population groups 2. Increasing awareness of schemes and benefits

25

2a. Awareness campaign by advertisements in media 2b. Awareness campaign targeted to specific places 3. Making the premium affordable

22

3b. Sliding-scale premium 4. Modifying enrolment

51

Strategy 2. Increasing awareness of schemes and their benefits Governments, or insurance schemes, took steps to make people aware of the health insurance schemes and their eligibility through the media or other channels. This was commonly reported (25 documents, three in LMICs), with two main categories.

4a. Simplifying enrolment procedure 4b. Integrating sources for enrolment 4c. Changing unit of enrolment 4d. Improving premium collection approaches 5. Improving health care delivery

16

5a. Improving health care package 5b. Controlling price of services 5c. Improving quality of services 6. Improving the management and organization

30

6a. Improving information system 6b. Staff training 6c. Transparent management

Strategy 1. Modifying the eligibility criteria This strategy included legislation or regulations to make uninsured populations eligible for health insurance schemes. This was the most commonly reported strategy from the USA, but only reported in two LMICs. There appeared to be two main categories. (a) Increasing the income threshold for entering health insurance. In order to enrol poor or low-income populations, many strategies described changing the income level for joining insurance scheme to increase accessibility to low-income groups. For example, in the USA, in California’s poverty expansion programmes for Medicaid, infants are eligible for Medicaid if their parents’ income is up to 200% of the Federal Poverty Line (FPL), children under 6 years old are eligible if it is up to 133%, for children 6–15 up to 100%, and for children 15 and older up to 83% of the FPL (Aizer 2002). In June 1998, the scheme expanded eligibility for public health insurance up to 250% of the FPL through its State Children Health Insurance Program (SCHIP) for children under 19 years old (Aizer 2002). (b) Expanding the categories of eligible population groups. Another approach is to change employee categories eligible for insurance. For example, in the USA, legal immigrants who have resided in the USA for more than 5 years are

(a) Mass media campaigns. In many countries, health insurance schemes are advertised to the potential eligible population via television, radio, print advertisement, internet or promotional hotlines. Some insurers even employ marketing experts or partner with other organizations to organize marketing campaigns. In California in the USA, for example, an advertising campaign via television, radio and print advertisements in 48 of California’s 58 counties was implemented in both English and Spanish (Aizer 2002). (b) Mass media at targeted locations. The SCHIPs in the USA sometimes implemented campaigns in venues where parents and children tend to congregate, including public benefit programmes, early childhood centres, schools, hospitals and religious institutions (Andrulis et al. 1999). Strategy 3. Making the premium affordable Use of subsidies or setting suitable premium levels to make insurance schemes affordable for eligible populations was found in 22 documents (six in LMICs). The two main categories of this strategy are as follows. (a) Subsidy. Government or other organizations directly or indirectly contribute all or part of the premiums for the eligible population, mainly involving three approaches: Government pays premiums for eligible populations. In the Philippines, the premium of the National Health Insurance Program (NHIP) for indigent households was covered by a government fund, the cost of the premiums being shared by the local government and the NHIP (Bautista et al. 1999). Tax credit to subsidize eligible populations indirectly. In Massachusetts in the USA, for example, the ‘health insurance connector’ facilitated the process of small employers’ enrolment by offering Section 125 plans, which allowed individuals to purchase health insurance plans using pre-tax dollars (Burton et al. 2007). Donation to pay premiums for the poor. In Kabutare (Rwanda), the local church paid for the contributions of about 3000 orphans and widows with their family members (Carrin et al. 2005).


3a. Subsidy

eligible for Medicaid and SCHIP on the same basis as US citizens; refugees and other humanitarian immigrants— except for parolees and domestic violence victims who are seeking legal status—are eligible for Medicaid and SCHIP regardless of their length of residence in the USA; and immigrants, both legal and undocumented, who meet all of the Medicaid eligibility requirements, except for the immigrant eligibility restrictions, can receive Emergency Medicaid if they need treatment for a medical emergency (Fremstad and Cox 2004).


(b) Sliding scale of premiums. Different premium levels are set based on the income level of the target population. In Bangladesh, for example, different contributions in the Gonosasthya Kendra scheme were set according to the social and economic status of the people (‘destitute’, ‘poor’, ‘middle class’ and ‘rich’) (Carrin et al. 2005). Strategy 4. Modifying enrolment Enrolment was maximized by improving the enrolment procedures. This category included 51 articles (eight in LMICs) with four mechanisms.

Strategy 5. Improving health care delivery This involved improving uptake by the population by covering a wider range of health care, controlling the price of services covered, or improving the quality of health care. This category was identified in 16 documents (eight in developing countries). (a) Improving the health care package. Insurance schemes met the needs of the eligible population by enlarging the health care package. In SCHIPs in the USA, children’s health insurance increases Child Health Plus benefits to include emergency, preventive and routine dental care, except orthodontics and cosmetic surgery; emergency, preventive and routine vision care, including eyeglasses; speech and hearing services; durable medical equipment; non-prescription drugs; outpatient mental health services and inpatient mental health, alcohol and substance abuse services (Andrulis et al. 1999). (b) Controlling the price of services. By adjusting the co-payment, deductible and/or ceiling, the services covered are made

affordable for the eligible population. In the USA, the amendments to Medicaid’s original rules in 1982 eliminated cost-sharing for children (Mann et al. 2003). (c) Improving the quality of services to attract more of the eligible population. In the district of Nouna (Burkina Faso), for example, most of the people took part in the community-based health insurance schemes with the expectation of improvements in quality of care which the health insurance management team promised during the enrolment campaign (De Allegri et al. 2006a). Strategy 6. Improving the management and organization of insurance programmes This category of strategies was included in 30 articles (14 in LMICs). (a) Improving the information system. Making use of an appropriate information system for measurement of the eligible population, enrolment and management of the schemes. In the development of SCHIP in the USA, the state of Maine improved data and information technology and Vermont created a state-wide integrated delivery system using an electronic database (Sasser 2007). (b) Staff training. Improving the ability of the management staff of the health insurance scheme for more effective outreach and running of the scheme. For example, in order to enrol Latino children, many states in the USA supported staff development and training in cultural competency to include appropriate family members and to deliver appropriate health care services (Zambrana and Carter-Pokras 2004). (c) Transparent management. Making the insured population engage in the design of insurance schemes. The prepayment schemes including benefit package, premium level, enrolment categories, co-payments and waiting period in Rwanda were designed, discussed and agreed upon (by voting) in a series of about 30 workshops in three districts, which were attended by the local population (Schneider and Diop 2005).

Location of the strategies Table 2 presents the distribution of strategies by region. All categories of strategies for expanding health insurance schemes were found in the literature from the USA. Most countries appear to strengthen the management of the insurance scheme (strategy 6; 16 countries), although this was often in combination with other strategies. Strategies adopted in LMICs focused on strategy 3 (making the premium affordable by subsidy or other methods), strategy 5 (improving health care delivery to attract populations) and strategy 6 (improving management and organization). Except for Bangladesh and India, where four and five categories of strategies were implemented for extending health insurance coverage, other LMICs adopted fewer than three categories of strategies. Target population of the strategies Rural and low-income populations were targeted in the majority of the included countries, especially in LMICs. In the USA, a wide range of strategies were adopted for the target population, while


(a) Simplifying the enrolment procedure. Here providers reduced the requirements for application, improved their forms or provided assistance to help applicants complete the forms. As an example, in California (USA) in 1998 the state government worked with a variety of organizations to provide application assistance to families who were potentially eligible for Healthy Families (California SCHIP) or Medicaid (Buchmueller and Jacobson 2007). (b) Integrating sources for enrolment. Insurance scheme partnered with other organizations or public programmes to facilitate enrolment. For example, in SCHIPs in the USA, children could be enrolled without the requirement of completing a new application form so long as another means-tested programme had already found that the family was poor or near-poor (Dorn 2007). (c) Changing the unit of enrolment. Some schemes changed to the family as unit of membership. An example was BadgerCare in Wisconsin, USA, where all family members were required to be enrolled (Gavin 2003). (d) Improving premium collection approaches. Changing the methods and timing of collecting premiums. For example, in the ORT Health Plus Scheme, a community-based health financing programme in the Philippines, premium collection was flexible, with monthly, quarterly or semi-annual payments. Burial societies in Uganda used their monthly meetings for the collection of premiums, either for first-time members or for those who renewed their membership (Carrin 2003).

5

6


Table 2 Location of the strategies from the studiesa Strategies

North America

Europe

South America

Asia

Modifying the eligibility criteria

USA (38)

France (1)

Haiti (2)

Egypt (1)

Africa

Increasing awareness of schemes and benefits

USA (22)

India (1)

Burkina Faso (1)

Making the premium affordable

USA (16)

India (1)

Rwanda (2)

Modifying enrolment

USA (43)

Germany (2)

Rwanda (1)

Philippines (2) Haiti (1)

Bangladesh (1)

Congo (1) Rwanda (2)

India (1)

Tanzania (1)

Philippines (2)

Uganda (2)

Improving health care delivery

USA (8)

Haiti (1)

Bangladesh (1)

Burkina Faso (1)

China (1)

Congo (1)

India (2) Philippines (1) Improving the management and organization

USA (11)

Belgium (1)

Brazil (1)

Bangladesh (1)

Congo (2)

France (1)

China (1)

Rwanda (2)

Germany (1)

India (2)

Tanzania (1)

Italy (1)

Philippines (1)

Uganda (1)

Netherlands (1)

Vietnam (1)

a

Figures in the table indicate the number of documents covering the strategies. In one article, more than one strategy could be found.

Table 3 Target population of strategies by country Target population

North America

Elderly

USA (2)

Children

USA (43)

Women

USA (5)

Rural population

Europe

South America

Germany (1)

Asia

Africa

Egypt (1) India (2)

Belgium (1)

Brazil (1)

Bangladesh (2)

Congo (1)

France (1)

China (1)

Rwanda (3)

Germany (1)

India (1)

Tanzania (1)

Italy (1)

Philippines (2)

Uganda (1)

Netherlands (1) Low-income population

USA (13)

France (1)

Migrants

USA (2)

Germany (1)

People with chronic diseases

USA (1)

Haiti (1)

India (1) Philippines (1)

in other countries only single or a limited number of strategies were implemented for those targeted (Table 3). For rural and low-income populations, studies in Bangladesh, India, the Philippines and Rwanda were found to use premium subsidies for expanding health insurance coverage.

2. What evaluation studies have been done and what have they shown? There were 25 studies evaluating strategies for expanding health insurance coverage. All the studies were conducted in

the USA and most of them focused on Medicaid, SCHIP and BadgerCare. Table 4 summarizes the characteristics of these studies.

Study design Most were time series analyses. Of the remainder, there was one controlled before-and-after study, one randomized controlled trial, one non-randomized controlled trial, two prospective intervention studies and one that used mixed methods, including interviews with policy makers.


Vietnam (1)

Modifying eligibility criteria

Associated strategies

Children with chronic health conditions

None

None

Time series

Time series

Time series

Expanding eligible criteria to Persons living people with HIV/AIDS with HIV or AIDS

Children in low-income families Children

Modifying eligibility standards through legislation

Law required Medicaid to cover all children up to age 6 with family income below 133% of poverty line by the year 2002

None

None

Time series

Low-income Use of Personal populations Responsibility and Work Opportunity Reconciliation Act to raise income-eligibility limits for adults

None

Retrospective cohort study

None

Low-income The programme extended women eligibility to a family income up to 185% of the FPL

Time series

Children

SCHIP was signed into law as part of the 1997 Balanced Budget Act

None

Time series

Children in near-poor families

Increased from 15% to 23%

Probability of being covered by public health insurance increased by 5.6–6.1% Raises the probability of public coverage by 8.1%

Medicaid

SCHIP

SCHIP

National Longitudinal Survey of Youth; 1989–1990, 1992– 1994

Panel data from Surveys of Income and Participation Program; 1986–1995

State HIV/AIDS Reporting and Surveillance system, etc.; 1992–1997

March Current Population Survey; 1994–2001

Lo Sasso and Buchmueller (2004)

Lee and Tian (2004)

Shore-Sheppard (2005)

Davidoff et al. (2005)

Increased maternal coverage by 7.7% and children by 8.9%

6–12% increase for HIV and AIDS coverage

Expansions resulted in increased Medicaid coverage Children without insurance fell from 15.5% to 9.1%

Medicaid

Medicaid

Medicaid

Medicaid

(continued)

Thorpe and Florence (1998–1999)

Ham and Shore-Sheppard (2001)

Bailey et al. (2003)

Aizer and Grogger (2003)

Increased the probability Wolfe et al. (2005) of having insurance by about 15%

A 9.8% increase in insured and a 6.4% decline in uninsured

SCHIP, Medicaid

Leininger (2007) Treatment group increased from 22% to 60%, and comparison group from 47% to 57%

Medicaid

Main author

Author’s main findings

Health plans

BadgerCare Administrative data of Wisconsin state; 1995, 1997, 1999

March Current Population Survey; 1996–2000

March Supplement of the Current Population Survey; 1996–2002

March Supplement Current Population Survey; 1988–1996

National Health Interview Survey; 1997, 2000, 2001

Yes

None

National Survey of America’s families; 1997, 1999, 2002

Data source and study period

None

Comparator or control

Time series

Controlled before-andafter study

Expanding the income eligibility thresholds to higher income children

Extending Medicaid coverage Children to families with incomes above the Aid to Families with Dependent Children (AFDC) thresholds

The State Children’s Health Insurance Program (SCHIP) expanded coverage to children

Poor adolescents Time series

Expand Medicaid-eligible to poor adolescents (ages 15–17)

Study design

Target population

Content of the strategies


Main strategies

Table 4 Studies assessing the strategies for expanding health insurance coverage


7

Awareness campaign

Main strategies

Time series

Time series

Children English and Spanish television advertisements; community-based application assistance Children Single short application; SCHIP premium exemption for American Indians/ Alaska Natives in 2001; expanding SCHIP coverage to childless adults up to 35% of the FPL in 2004

Modifying enrolment

Making premium affordable; modifying enrolment

None

None

None

Document analysis, and qualitative interviews

Low-income children

Distribute information and provide application assistance through various public agencies; coordinate different schemes

None


Low-income Provide assistance with elisingle gibility and application mothers questions over a telephone hotline, require no face-to-face interview, etc.

Modifying enrolment

None

Non-randomized Yes controlled trial

None


Children

Time series

Modifying eligibility standChildren ards; awareness campaign

Handing out insurance applications in the emergency department

Advertisement of the health Children insurance scheme in collaboration with other organizations; use of community-based organizations to facilitate application procedure; infants are eligible for medical cover up to 200% of the FPL

Awareness campaign; modifying enrolment

None

Time series

Children; women; low-income populations; populations with disability

Expand eligible populations; use of tax and voluntary donation programmes to support the poor

Making premium affordable

None

Time series

Elderly (poor seniors)

Expanding insurance eligibility criteria to the elderly; pay for premium and cost-sharing for the poor elderly


Comparator or control

Study design

Target population

Content of the strategies

Associated strategies

More poor populations were enrolled into insurance

Medicaid

Medicaid

Medicaid Statistical Information System; 1999–2003

MIChild (SCHIP in Michigan)

Coverage raised by 45%

Advertising increased enrolment by 4.7%

(continued)

Ellis et al. (2007)

Aizer (2006)

Dunbar et al. (1999)

Wolfe and Scrivner (2005)

Increased by 6–25%

Five states enrolled over 13 000 new enrollees

Cullen et al. (2005)

4.3% participation rate increased

SCHIP; Medicaid

Gordon et al. (2005)

14% increase after intervention

SCHIP

Medicaid enrolment data Medicaid and administrative data; 1996–2000

State documents; 1997– 1999

Adams (1995)

Yelowitz (1997)

Main author

Enrolment rates increase Aizer (2002) in months after an advertisement

Insured rate increased from 8% to 12.5%

Medicaid

SCHIP

Author’s main findings

Health plans

Wisconsin Administrative data BadgerCare from the State of Program Wisconsin; 1995–1997

Early Childhood Longitudinal Study; 1998–2002

Survey data in randomized controlled trial (RCT); 2001–2002

Current Population Survey; 1996–1999

Statistical report on Medical Care, etc.; 1984–1991

Survey of Income and Program Participation; 1987, 1988, 1990, 1991, 1992, 1993

Data source and study period


Table 4 Continued

8 HEALTH POLICY AND PLANNING

Bansak and Raphael (2005) SCHIP March Current Population Survey; 1997–2001

A range of 7.6–17.1% increased by different methods

31% could be traced dir- Gordon and Dupuie ectly to ED-based (2001) intervention efforts Prospective observational SCHIP study for the 12 month period; 1998– 1999

Medicaid application increased 22% Medicaid Hospital registration data; 2001–2002

Mahajan et al. (2005)

Intervention group had 8 Flores et al. times the adjusted (2005) odds of having insurance. SCHIP, Medicaid Telephone interview (RCT); 2002–2004

Increased probability of coverage by 3.1% Private health insurance National Longitudinal Survey of Youth; 1989–1998

Baughman (2001)

Gumus and 0.68% and 1.02% Regan (2007) increased by 10% decrease in the after-tax price Private health insurance Current Population Survey; 1995–2005

None

None

Retrospective observational study

Repeated crosssectional study

Children

Children in low-income families

Enrolment assistance in emergency department

Eliminated Asset Test; 12-month continuous eligibility

Modifying enrolment


Retrospective observational study

None Children Emergency department provides application assistant

RCT

Yes Latino children Enrolment by the case managers to insure uninsured Latino children

Retrospective cohort study Low-income workers Earned Income Tax Credit provides refundable tax credits

None

Time series Self-employed population Deducting 25% of health insurance premium from taxable income

None

Main author Author’s main findings Data source and study period Comparator or control Study design Target population Content of the strategies

Data sources and analysis Most data sources were surveys or cohorts using existing databases. For the time series, econometric analysis was the main analytical approach. All evaluations reported on expansion of coverage (Table 4). Follow-up was generally between 3 and 5 years. Strategies Predominantly these changed the population groups eligible by increasing the income threshold for entering health insurance. In addition, awareness campaigns, offering subsidies to low-income people and modifying the enrolment approaches were the major strategies found from the studies looking at coverage expansion. Outcomes The main outcome indicators include changes in coverage of the health insurance schemes or in probability of the study sample being insured by the interventions. As assessed by the authors, all the studies found that the interventions studied were positive in expanding the coverage of the health plans.

Discussion

Associated strategies Main strategies

Table 4 Continued

9

In this paper we have synthesized across many papers, using the framework approach, strategies that have been used to expand or improve health insurance plans or community-based health insurance schemes to vulnerable groups. The majority of the literature comes from the USA; studies from LMICs were limited. Strategies fell into six basic categories, and many schemes used more than one strategy when they set about expanding coverage. This is to be expected, that strategies to expand health insurance coverage within a region or country were likely to be a set of individually complex interventions (Petticrew 2009), and means that findings around the success or otherwise are likely to be specific to the mix of strategies, and to the specific context in which they were designed. This presents a difficulty in trying to apply evidence-based decision-making based on reviews of countries or programmes that have used experimental or quasi-experimental methods. The nature of this review is a descriptive scoping review that aimed to systematically identify the strategies implemented for expanding health insurance coverage, with an additional summary of evaluation studies. The methods for inclusion of eligible documents in this review meant that documents available that described or evaluated the strategies actually implemented were included. However, we will only detect a strategy if it has been reported on in the grey or published literature. The value of this approach is that it provides a summary of the strategies that have been used for expanding health insurance coverage. This provides policy makers with a series of options. Second, this mapping can then prioritize further reviews either on the effects of strategies or more detailed descriptive reviews. From all the included studies, evaluation documents could be picked up in an effect review examining the effectiveness of interventions for expanding health insurance coverage. Or, more detailed reviews could be conducted on


Health plans


10


Health systems research objectives and methods are usually not suitable for effect reviews. In addition, health systems research usually targets a package of interventions and closely links interventions with complicated contexts. These challenges make it difficult to select interventions, study types and study inclusion criteria for an evaluative systematic review. There needs to be further methodological development in drawing out the results from one research study in one context and applying it somewhere else. This review tries addressing these challenges by conducting a descriptive systematic review of options for policy but not going further into formal effectiveness analysis; this was not relevant to the questions we posed.

Acknowledgements The team thank Dr Sara Bennett, the Manager of the AHPSR, for providing coordination and technical supports; thanks to Prof. Diane McIntyre from the University of Cape Town and Dr Kara Hanson from the London School of Hygiene and Tropical Medicine for their contributions in developing the review protocol; thanks to Dr Kent Ranson from the AHPSR and Dr Sandy Oliver from the University of London for their contributions to the development of the protocol and comments on the draft. We are grateful to Ms Li Wenjun, Mr Pang Guohua and Mr Liu Bin from the Center for Health Management and Policy, Shandong University, for their help in collection of the literature.

Funding This work was financially supported by the Alliance for Health Policy and Systems Research, the World Health Organization.

Authors’ contributions M.Q. contributed to design, methods, review and writing; B.B.Y. contributed to methods, review and writing; L.J. contributed to review and writing; J.W. contributed to methods and review; B.Y. contributed to methods and review; G.P. contributed to design and writing.

Conflict of interest The authors declare that they have no competing interests.

References Adams EK. 1995. Equity in the Medicaid program: changes in the latter 1980s. Health Care Financing Review 16: 55–73. Aizer A. 2002. The use and effect of public programs on child health and well-being. Thesis, University of California, Los Angeles, USA. Aizer A. 2006. Public health insurance, program take-up, and child health. NBER Working Paper. Cambridge, MA: National Bureau of Economic Research.


specific aspects, e.g. focusing on low-income countries or focusing on social health insurance schemes. The review objectives and methods was not intended to provide a rigorous assessment of the effectiveness of the strategies implemented. This review is an initial step in considering which strategies may be worth evaluating, through review or primary research, in much more depth. This potentially could reduce the value of the review in informing policy makers and academics on what strategies really work. One finding that is important is that in expanding coverage, and in evaluating strategies that do so, we cannot target single interventions, because most of the included documents present a combination of strategies, which is what policy makers will undertake in practice. There are questions about the appropriateness of effects review methods (such as Cochrane) in health systems development (Pawson et al. 2005). We knew from the outset, from the existing systematic reviews of effects, that there were no rigorous quasi-experimental or experimental studies that had evaluated any strategy in expanding coverage. To carry out a Cochrane review would add little and provide a framework for ‘evidential nihilism’ (Petticrew 2009), which does not help take the debate forward. The Cochrane review on this topic, which we carried out after the scoping study, includes only two small studies from the USA (Meng et al. 2010). Even though we cannot make generic conclusions about the policy implications of strategies in different settings, policy making in expanding health insurance coverage in one context could be stimulated by practice in other contexts. For example, many of the strategies found in the USA for expanding coverage are non-financial, implying that these strategies are not expensive to implement and thus are ones from which LMICs could learn. China’s rural health insurance scheme demonstrates a good case in point. Its rapid expansion (91.5% of the rural population had been covered by the end of 2008, just 5 years after its inception in 2003) benefited much from education and other non-financial activities to the people, aside from an increase in government subsidies (Li 2008; Ministry of Health, China 2009). Another value of this review in relation to policy making is that the description of strategies for expanding health insurance coverage in LMICs provides a range of options that countries with similar contexts could consider. As LMICs move forward with programmes of insurance, it would be worthwhile attempting to generate some standard approaches to describing the various sets of complex interventions that they may adopt. In addition, we need to generate a set of generic context factors to describe the country, health status and people that will help to explain the strategies used, and also the potential of the context to influence whether the strategies are feasible, appropriate and effective. Where possible, in the process of implementing these strategies, quasi-randomized methods could be nested into the implementation package to explore the effectiveness of different approaches. For example, the income threshold could be altered, and this monitored by coverage before and after; and at the same time, different approaches to simplifying enrolment could be evaluated at different sites using cluster randomized or controlled before-and-after design.


Aizer A, Grogger J. 2003. Parental Medicaid expansions and health insurance coverage. NBER Working Paper. Cambridge, MA: National Bureau of Economic Research. Andrulis DP, Bauer TA, Hopkins S. 1999. Strategies to increase enrolment in children’s health insurance programs: a guide to outreach, marketing and enrolment in New York and other states. A report of the New York Academy of Medicine. Journal of Urban Health 76: 247–79. Arhin-Tenkorang D. 2001. Health insurance for the informal sector in Africa: design features, risk protection, and resource mobilization. HNP Discussion Paper. Washington, DC: World Bank.

Bansak C, Raphael S. 2005. The effects of state policy design features on take up and crowd out rates for the State Children’s Health Insurance Program. Working paper. San Diego, CA: Department of Economics, Center for Public Economics, San Diego State University. Baughman RA. 2001. Three essays on the behavioral impacts of public policy on health. Thesis. Syracuse University, NY, USA. Bautista MC, Yap EC, Soriano ES. 1999. Local governments’ health financing initiatives: evaluation, synthesis, and prospects for the national health insurance program in the Philippines. Report. Bethesda, MD: Partnerships for Health Reform (PHR).

Dunbar JL, Sloane HI, Mueller CD. 1999. Implementation of the State Children’s Health Insurance Program: outreach, enrolment, and provider participation in rural areas. Report. Bethesda, MD: Project HOPE Walsh Center for Rural Health Analysis, and Rockville, MD: Office of Rural Health Policy, Health Resources and Services Administration. Ekman B. 2004. Community-based health insurance in low-income countries: a systematic review of the evidence. Health Policy and Planning 19: 249–70. Ellis E, Walls J, Marks D, Trenholm C. 2007. Covering kids & families: case study in Michigan: Exploring Medicaid and SCHIP enrolment trends and their links to policy and practice. Report. Princeton, NJ: Robert Wood Johnson Foundation. EPPI-Centre. 2008. EPPI-Reviewer software. London: EPPI Centre, Social Science Research Unit, Institute of Education. Online at: http://eppi.ioe.ac.uk/cms/Default.aspx?tabid¼1913&language ¼en-US. Flores G, Abreu M, Chaisson CE et al. 2005. A randomized, controlled trial of the effectiveness of community-based case management in insuring uninsured Latino children. Pediatrics 116: 1433–41. Fremstad S, Cox L. 2004. Covering new Americans: a review of federal and state policies related to immigrants’ eligibility and access to publicly funded health insurance. Report. Menlo Park, CA: Kaiser Family Foundation. Gavin N. 2003. Evaluation of the BadgerCare Medicaid Demonstration. Report. Research Triangle Park, NC: Research Triangle Institute.

Buchmueller TC, Jacobson M. 2007. Can private companies contribute to public outreach efforts? Evidence from California. Health Affairs 26: 538–48.

Gilson L, Kalyalya D, Kuchler F et al. 2000. The equity impacts of community financing activities in three African countries. International Journal of Health Planning and Management 15: 291–317.

Burton A, Friedenzohn I, Vidal EM. 2007. State strategies to expand health insurance coverage: trends and lessons for policymakers. Bipartisan Congressional Health Policy Conference, 12–14 January, Miami, Florida.

Gordon JA, Dupuie TA. 2001. Child health insurance outreach through the emergency department: a pilot study. Academic Emergency Medicine 8: 1088–90.

Carrin G. 2003. Community based health insurance schemes in developing countries: facts, problems and perspectives. Discussion Paper, EIP/FER/DP.E.03. Geneva: World Health Organization. Carrin G, James C. 2005. Social health insurance: key factors affecting the transition towards universal coverage. Internal Social Security Review 58: 45–64. Carrin G, Waelkens MP, Criel B. 2005. Community-based health insurance in developing countries: a study of its contribution to the performance of health financing systems. Tropical Medicine and International Health 10: 799–811. Cullen JB, DeCicca P, Volden C. 2005. The impact of state CHIP programs on early childhood health insurance coverage, utilization and outcomes. ERIU Working Paper 52. Ann Arbor, MI: Economic Research Initiative on the Uninsured, University of Michigan. Davidoff A, Kenney G, Dubay L. 2005. Effects of the State Children’s Health Insurance Program expansions on children with chronic health conditions. Pediatrics 116: 34–42. De Allegri M, Sanon M, Bridges J, Sauerborn R. 2006a. Understanding consumers’ preferences and decision to enroll in community-based health insurance in rural West Africa. Health Policy 76: 58–71. De Allegri M, Kouyate´ B, Becher H, Gbangou A. 2006b. Understanding enrolment in community health insurance in sub-Saharan Africa: a population-based case-control study in rural Burkina Faso. Bulletin of the World Health Organization 84: 852–8.

Gordon JA, Emond JA, Camargo CA. 2005. The state Children’s Health Insurance Program: a multicenter trial of outreach through the Emergency Department. American Journal of Public Health 95: 250–3. Gumus G, Regan TL. 2007. Tax incentives as a solution to the uninsured: Evidence from the self-employed. IZA Discussion Paper No. 2866. Bonn, Germany: IZA (Institute for the Study of Labour). Habtom GK, Ruys P. 2007. Traditional risk-sharing arrangements and informal social insurance in Eritrea. Health Policy 80: 218–35. Ham JC, Shore-Sheppard LD. 2001. The effect of Medicaid expansions for low-income children on Medicaid participation and insurance coverage: evidence from the SIPP. NBER Working Paper. Cambridge, MA: National Bureau of Economic Research. Hsiao W, Shaw P. 2007. Social Health Insurance for Developing Nations. Washington, DC: World Bank. Lagarde M. 2006. Evidence from systematic reviews to inform decision making regarding financing mechanisms that improve access to health services for poor people. A policy brief prepared for the International Dialogue on Evidence-Informed Action to Achieve Health Goals in Developing Countries (IDEAHealth). Geneva: Alliance for Health Policy and Systematic Research, World Health Organization. Lee HJ, Tian WH. 2004. The State Children’s Health Insurance Program: Participation and Substitution. ERIU Working Paper 53. Ann Arbor, MI: Economic Research Initiative on the Uninsured, University of Michigan.

Devadasan A, Ranson K, van Damme W, Acharya A, Criel B. 2006. The landscape of community health insurance in India: an overview based on 10 case studies. Health Policy 78: 224–34.

Leininger L. 2007. The impact of the Medicaid expansions of the late 1990’s on the insurance coverage of poor adolescents. Conference draft. Ann Arbor, MI: Economic Research Initiative on the Uninsured, University of Michigan.

Dorn S. 2007. Eligible but not enrolled: how SCHIP reauthorization can help. Issue brief. Washington, DC: Urban Institute.

Li XY. 2008. A study on NCMS coverage in Beijing. Southern Rural 1: 30–3.


Bailey JE, Van Brunt DL, Raffanti SP, Long WJ, Jenkins PH. 2003. Improvements in access to care for HIV and AIDS in a statewide Medicaid managed care system. American Journal of Managed Care 9: 595–602.

11

12


Lo Sasso AT, Buchmueller TC. 2004. The effect of the state children’s health insurance program on health insurance coverage. Journal of Health Economics 23: 1059–82. Mahajan P, Stanley R, Ross KW et al. 2005. Evaluation of an emergency department-based enrolment program for uninsured children. Annals of Emergency Medicine 45: 245–50. Mann C, Rowland D, Garfield R. 2003. Historical overview of children’s health care coverage. Future of Children 13: 31–53. Meng Q, Yuan B, Jia L, Garner P. 2010. Outreach strategies for expanding health insurance coverage in children. Cochrane Database of Systematic Review. Issue 8, Art. No. CD008194. DOI:10.1002/ 14651858.CD008194.pub2. Ministry of Health, China. 2009. Health systems development in China 2008. Beijing: Ministry of Health.

Petticrew M. 2009. Systematic reviews in public health: old chestnuts and new challenges. Bulletin of the World Health Organization 87: 163. Sasser A. 2007. Reaching the goal: expanding health insurance coverage in New England: current strategies and new initiatives. NEPPC Research Report no. 07–1. Boston, MA: New England Public Policy Center, Federal Reserve Bank of Boston. Schneider P, Diop F. 2005. Community-based health insurance in Rwanda. In: Preker AS, Carrin G (eds). Health Financing for Poor

Shore-Sheppard LD. 2005. Steming the tide? The effect of expanding Medicaid eligibility on health insurance. NBER Working Paper. Cambridge, MA: National Bureau of Economic Research. Thorpe KE, Florence CS. 1998–1999. Health insurance among children: the role of expanded Medicaid coverage. Inquiry 35: 369–79. WHO. (2005). Sustainable health financing, universal coverage and social health insurance. 57th World Health Assembly, Ninth plenary meeting, 25 May 2005, Resolution WHA58.33. Wolfe B, Scrivner S. 2005. The devil may be in the details: how the characteristics of SCHIP programs affect take-up. Journal of Policy Analysis and Management 24: 499–522. Wolfe B, Kaplan T, Haveman R, Cho YY. 2005. Extending health care coverage to the low-income population: he influence of the Wisconsin BadgerCare Program on insurance coverage. IZA Discussion Paper no. 1546. Bonn, Germany: IZA (Institute for the Study of Labour). Yelowitz AS. 1997. Public policy and health care choices of the elderly: evidence from the Medicare Buy-In Program. IRP Discussion Paper no. 1136–97. Madison, WI: Institute for Research on Poverty, University of Wisconsin-Madison. Zambrana RE, Carter-Pokras O. 2004. Improving health insurance coverage for Latino children: a review of barriers, challenges and State strategies. Journal of the National Medical Association 96: 508–23.


Pawson R, Greenhalgh T, Harvey G, Walshe K. 2005. Realist review – a new method of systematic review designed for complex policy interventions. Journal of Health Services Research and Policy 10: 1–34.

People: Resource Mobilization and Risk Sharing. Washington, DC: World Bank, pp. 251–74.