The Chlamydia trachomatis (Ct) bacterium is the most frequently reported infectious disease in the United States.1 Left untreated, Ct may lead to ectopic pregnancy, infertility, pelvic inflammatory disease, and an increased susceptibility to human immunodeficiency virus infection. Owing to the health risk associated with Ct, national guidelines set by the US Preventive Task Force and the Centers for Disease Control and Prevention (CDC) recommend that sexually active females younger than 25 years are routinely screened for Ct.2,3
Although there are established national guidelines for Ct screening, studies have demonstrated inadequate risk assessment and education about the infection. A recent survey found that less than half of young sexually active women are receiving routine Ct screening in managed care organizations, although screening rates have been increasing.4 C. trachomatis is a prime candidate for disease prevention because it is highly prevalent, associated with high morbidity, and able to be diagnosed and treated. C. trachomatis screening is essential because it allows for prevention and treatment of the infection, if detected. In addition, Ct screening will prevent pelvic inflammatory disease and other related complications.
This study presents findings from an evaluation of Ct screening activities implemented in commercial health plans. The purposes of this study are (a) to ascertain what activities commercial health plans have attempted (if any) to increase their Ct screening rates and (b) to identify barriers to improving Ct screening rates in commercial health plans. C. trachomatis screening rates are considerably lower than the rates for other similar screening tests; for example, 38% of eligible women receive a Ct screening test, whereas 83% of eligible women received a screening test for cervical cancer.5,6
MATERIALS AND METHODS
We used a qualitative study design to explore the Chlamydia screening practices of commercial health plans. Structured interviews were conducted with health plan contacts knowledgeable about Ct screening activities and/or in charge of overseeing Ct screening policies and procedures.
Health plans were identified by the research team using the Ct screening rate provided in the National Committee for Quality Assurance (NCQA) Healthcare Effectiveness Data and Information Set (HEDIS) data.7 National Committee for Quality Assurance’s Ct HEDIS measure reports the percentage of sexually active female plan members who had at least 1 test for Ct during the measurement year and is currently reported separately for women aged 16 to 20 and 21 to 24 years. The denominator for the Ct HEDIS measure is the proportion of women in the age range who (a) sought care and (b) were determined to be sexually active. The numerator includes the women in the denominator who were screened for Ct. This measure has been found to be highly correlated with actual screening rates in previous work.8 The NCQA audits the HEDIS data with certified auditors to ensure the validity of the data.
To identify plans to interview, plans were arrayed by screening rate (to identify low and high performers) and then by changes (to identify plans with recent declines or increases in Ct screening rates). The data set contained HEDIS scores, for a total of 394 reported plans. However, there were fewer independent plans. For example, one national health plan appeared 65 times in the data—once for each state they operated in with either an HMO or a PPO plan.
To select plans for interview, we first applied exclusion criteria then arrayed the plans in order of screening rates. The exclusion criteria include if the plan uses a staff model and if the plan had less than 500 eligible women. We then selected plans beginning with the highest (lowest) screening rate then proceeding up (down) the list until we had interviewed our target number of plans. However, an attempt was made, to the extent possible, to select plans from different regions of the country. This was done by selecting the highest (lowest) performing plans both from the overall list and at least one plan from each of the region-specific lists.
After identification, NCQA contacted each plan individually asking the plan to participate in the research program. Health plan quality assurance and/or data managers were contacted via e-mail by an NCQA staff member who informed them about the purpose of the study and invited them to participate. The research team then sent a follow-up e-mail to health plans to inquire about their interest in participating. If a response was not received, researchers telephoned the health plan staff member to ask the plan to participate in the research.
All interviewed plans were fairly large—this was to ensure that the plans had sufficient enrollment among eligible women. The number of members ranged from a high of 2.5 million to a low of 26,000. The number of eligible women ranged from a high of more than 20,000 to a low of 1100.
We developed and pilot tested 4 related qualitative interview questionnaires. Each protocol consisted of a series of open-ended questions with accompanying probes designed to elicit information from health plan officials about their plan’s Chlamydia screening rates. Each interview protocol emphasized the broad themes of clinical practice guidelines for Chlamydia, use of interventions to increase Ct screening rates among plan providers, facilitators and barriers to these interventions, member and provider health education specific to Chlamydia screening, and, finally, collection of Chlamydia screening data. In addition, the questionnaires for low-performing plans focused more extensively on barriers to implementation.
The initial questionnaires were developed by the research team in consultation with CDC staff. These questionnaires were then vetted with an expert review panel, which made additional suggestions for areas of emphasis and questions that were either considered appropriate or inappropriate. After finalizing the format, we conducted 3 pilot interviews for both high- and low-performing plans. The pilot interviews confirmed the appropriateness of the instruments. Each interview was conducted by telephone and lasted between 45 minutes to 1 hour. Interviews were audiorecorded with the permission of the participant. An informed consent form was orally read to each participant, after which the participant was asked to consent to the interview. The consents were then audiorecorded and noted in our files. The study received appropriate human subject approval from Emory University’s institutional review board. Participation in this study was voluntary, and all data were kept confidential.
Interviews were professionally transcribed. Interviewers also kept notes from each interview that were used in combination with the transcriptions to produce an interview summary that captured the Ct screening activities of each health plan. Content from both the low-performing health plans and high-performing health plans was analyzed according to the interview questions and organized into frameworks of categories and subcategories determined by the responses. Key themes were then placed into summary matrixes to facilitate analysis. Both interviewers reviewed all summaries and matrices and found consensus with the key themes that were identified.
The overall response rate was 64% (35 interviews from 55 contacts). In the first round of interviews (for high-performing and improving plans), a total of 36 plans were contacted and 22 interviews were completed. In the second round of interviews (for low-performing and declining plans), a total of 19 plans were contacted and 13 interviews were completed.
Figure 1 shows the distribution of screening rates in 2007. Although there were a small number of successful plans, most plans were relatively unsuccessful in getting most of their eligible members to receive a Ct screening test. Although there is a wide gap between the lowest plan (16%) and the highest plan (92%), the gap between the 25th and 75th percentile is only 11%—ranging from 30% to 41% (Fig. 1). Most plans screened between 25% and 50% of their eligible members.
There was no clear relationship between plan size and screening rates. Figure 2 shows the percentage of eligible members successfully screened ordered by plan size, from the smallest (38 eligible members and a 92.1% screening rate in 2007) to the largest (83,540 eligible members and a 63.5% screening rate). The relationship is mildly positive, with larger plans having slightly higher screening rates. However, the relationship is not strong and is not linear.
There were a number of commonalities across all the interviewed plans. First, all of the plans, with one exception, reimbursed providers using a fee-for-service methodology whereby the provider was directly paid for services provided. The exception used capitation, whereby the provider was paid a fixed amount per month for all services provided. However, a number of plans reported that they used other reimbursement methodologies in addition to the standard fee-for-service payment methodology, such per-diem or capitation payments. For example, most of the plans reported that the plan operated some type of pay-for-performance (P4P) program. Among the low-performing plans, 9 of 13 reported P4P programs and 21 of the 22 high-performing plans had such programs. However, none of the plans reported including Ct screening in their P4P plans. The P4P programs instead focused on conditions such as diabetes, cardiac diseases, and asthma. One plan commented “this is a decision of the medical director and our Quality and Utilization Committee. The physicians on that committee felt that the items noted above (e.g., diabetes) are of a greater priority based on our population data with the most direct impact on medical costs. Of course, priorities may change.”
Second, all plans reported that providers were specifically reimbursed for Ct screening. Furthermore, no plans reported complaints from providers about reimbursement levels. Several plans mentioned that they do hear from providers periodically about reimbursement issues but that Ct screening had not been one brought up by providers.
Third, every plan interviewed reported using practice guidelines. Most plans reported using nationally developed guidelines; the US Preventive Task Force guidelines, NCQA, and American Congress of Obstetricians and Gynecologists were specially mentioned by several plans. Other plans reported that guidelines were drawn from either an in-house committee—such as a Physician Advisory Committee—or a state or regional collaborative. All plans but one reported that Ct screening is included in their guidelines. Several plans reported that providers were resistant to plan provided guidelines; however, others reported that providers accepted the guidelines because they mirrored recommendations from nationally accepted groups. However, most plans expressed skepticism about the effectiveness of guidelines. No plan reported aggressive efforts to induce providers to follow Ct screening guidelines, and few reported efforts to induce providers to follow any guidelines, partly due to the inclusion of girls younger than 18 years in the screening guidelines. This is exemplified by a health plan that reported that they “always experience push-back on the screening guidelines because of the age group.”
Specific Ct Screening Interventions
Of the 35 plans interviewed, only 3 plans reported no Ct-specific efforts. The other plans had a variety of interventions designed to improve their Ct screening rate. These interventions included the following:
* Newsletters targeting Ct screening rates
* Providing providers with a Ct fact sheet
* Letters from the medical directors to providers detailing the provider’s Ct screening rate, description of the guidelines, and suggestions for ways to improve screening rates
* Development of an adult preventive health flow sheet as a best-practice activity, with Ct screening included
* “Patient profiles” sent to providers, which list patients who have “fallen out” and did not receive recommended screenings
* “Birthday letters” sent to plan members listing needed screenings tests (including Ct screening)
* “Physician profiles” listing more than 20 clinical measures—including Ct screening—and how the provider is doing relative to peers
* Meetings with ob-gyn groups by the medical director to show the group’s Ct screening rates and to discuss strategies to improve the screening rate
* Similar meetings with individual providers (there were different approaches to this—one plan used a prenatal care coordinator who knew the providers personally, other plans used medical directors of quality improvement staff, and one plan developed a slide show presentation that was shown to 1000 of the plan’s 3000 providers)
* One plan ordering automatic Ct screening (when the laboratory receives an order for a Papanicolaou test for a woman in the target age, the plan has the laboratory automatically perform a Ct screen and send the result to the provider)
Many of the interventions listed were used by multiple plans (particularly newsletters and birthday cards). Overall, the interviewed health plans were—with a small number of exceptions—very aware of the issue and their low screening rates and were actively engaged in improving the Ct screening rates. This was equally true of high- and low-performing plans.
Only 5 of the 35 interviewed plans reported no patient or provider educational efforts. All other plans reported active efforts to educate patients, providers, or both. Most of these were relatively low-key efforts such as the (already mentioned) birthday cards, sending out plan guidelines to providers or patients, or newsletters.
One plan offered that, from their perspective, mailings to patients are “…very high level but just kind of gets to the point of why it’s something you might want to think about if you’re potentially a candidate that could get that disease.” Generally, on direct mailings, the plan’s medical director signs the letter and suggests that each member talk to her physician if she thinks screening is appropriate. The goal of the letter is to reach out to members to raise awareness. The plan has not received any negative feedback on the mailing; however, the plan only contacts women older than 19 years because of concerns about negative feedback if younger women were contacted. For younger members, the plan is now sending out a brochure that discusses different risky behaviors, with a discussion of Chlamydia “buried in there,” among other issues. The goal of the brochure is to stimulate thinking about issues such as alcohol and drug abuse.
Perceived Barriers to Improving Ct Screening Rates
All plans but one identified important barriers that the plan faces in trying to improve their Ct screening rates. These barriers generally fell around 6 key issues:
1. Data collection issues
2. Identifying sexually active members
3. Appropriateness of the use of screening test
4. Limited resources
5. Problems with minors
6. Members receiving Ct screenings from outside providers
The first problem, data collection issues, reflect problems plans faced in capturing accurate Ct screening data for health plans. One plan commented that because providers only submit encounter data, this established “a new level of complexity” for the plan’s data collection activities. Some plans required physicians to use laboratories selected by the health plan and received Ct screening results from the laboratory concurrently with the physician. Other plans use Logical Observation Identifiers Names and Codes (LOINC). Logical Observation Identifiers Names and Codes are universal identifiers for laboratories that standardize the coding for laboratory testing. Using LOINC codes for services such as Ct screenings facilitates the electronic capture of screening data.
However, most plans reported capturing Ct screening from current procedural terminology (CPT) codes. However, there was considerable variability in how the data were collected. Different approaches included the following:
* Entirely collected from CPT or other administrative claims data
* Provided by an outside vendor that “scrubs” the administrative data and provides a report to the plan
* Reports directly from laboratories owned by the plan or contracting directly with the plan
* Reports from laboratories through providers who own or contract with laboratories
Most plans reported combinations of these different approaches. Several plans expressed skepticism about the quality of the data. Particular problems were with capitated providers and laboratories. Different vendors would also used different coding approaches; for example, the plan might contract with a national laboratory that uses LOINC codes, but some providers would use a local vendor without the LOINC codes. The plan would therefore have to use a different approach to calculate the screening rates for the providers using the national laboratories—the LOINC codes—and the providers who did not. For those providers, the plan might use claims codes, administrative data, or a combination. Plans also reported that providers were inconsistent in how they coded Ct screening with CPT codes. Plans reported that providers might use up to 4 different coding schemas.
The second problem, identifying sexually active members, revolves around the issue of the accuracy of the HEDIS Ct screening measure and whether the reported screening rate accurately measures the true rate of appropriate screenings. The plans had 2 key complaints. First, plans questioned the accuracy of the algorithm used to determine who is sexually active. Plans reported that providers often prescribe birth control pills for reasons other than sexual activity. The plans did believe that pregnancy tests were valid measures of sexual activity but that birth control pill subscriptions were not. One plan cited an NCQA study showing a 21% error rate in determining who is sexually active. The plans thus argued that the denominator is too big, making their screening rates too low. Another plan commented “it’s hard to determine accurately if women are sexually active and who should be screened.” A second related issue is that younger women often do not come to the doctor at all, so achieving a Ct screen for younger women is impossible.
We labeled the third barrier as appropriateness of the use of screening test. In many cases, the providers believed that there may have been mitigating circumstances that made the test inappropriate. For example, a married woman in a monogamous relationship may not be at risk and therefore not tested. A plan reported that providers “don’t want to explain to a married woman why they have to be tested for a sexually transmitted disease.” One plan commented that if the plan sent materials to the provider identifying a woman who should be screened for Ct and the woman was not sexually active, it would harm the plan’s credibility for other measures. One plan offered that they had experienced a “fair amount of push-back from providers about Chlamydia screening.”
The fourth barrier mentioned by many plans was limited resources. There were a number of variants on this issue. Plans reported that they themselves had insufficient resources to target everything that was important and that the plan did not feel that Ct screening was important enough to take resources away from diabetes or heart disease (as examples). Others reported that they had limited resources and targeted conditions that employers cared about, which did not included Ct screening. Finally, the plans feel that they have a limited ability to influence providers and that Ct screening is not an area where the plan wants to use up their influence, particularly because providers are already doubtful about the measure itself. One plan commented that:
I hate to say this, but in terms of prioritizing things that are important to us as a health plan, Chlamydia hasn’t made the short list, if you will. It’s not something that is awarded any points related to your performance on that measure for NCQA’s accreditation. And I think that is the reason it has been prioritized very low.
However, Ct screening is now part of NCQA accreditation scores, so this may have changed since the time of the interview.
The fifth barrier was issues with minors. Most plans reported some variant on this concern. Plans had issues about how to encourage minors to come to providers. More important were potential clashes with the moral values of the patents if the plan is perceived as suggesting that a minor is sexually active, encouraging a minor to be sexually active, or even facilitating sexual activity among minors. Plans feel that providers feel that this is potentially inappropriately interference with a family relationship. Another issue is the potential problem if a Ct screening code shows up on a bill sent to parents who were unaware their minor child was sexually active. One plan commented that they “don’t touch the issue with young girls,” whereas another said that “the culture in the state is focused on abstinence. This is changing over time, but the providers are not changing as rapidly as the teenagers.” Another plan reported that “I don’t think people want us to be that intrusive, especially since it gets down to the 16 and 17 years old. It’s an awkward issue for us...we don’t want to be in the newspaper about THAT…about that way…‘they’re sending letters to my daughter’.” Another plan lamented “…the lack of ability to do a real direct consumer (campaign) …given that the measures are for underage kids….its 16, 17, and 18 year olds and how do you message …well, sexually activity without condoning sexual activity.” This was a challenge to them because the health plan has been successful in developing “very crisp, straightforward member education materials” in other areas.
The sixth issue mentioned by plans is the belief that some members may actually receive Ct screenings from providers not captured by the health plan’s data. The plans suggest that women may receive Ct screening from organizations such as a public health department or Planned Parenthood. These organizations may offer subsidized or free screenings in a more anonymous environment than that provided by the health plan. One plan commented that “theres a reason all the planned parenthoods are located near college campuses.”
The final issue mentioned by a number of plans is culture. Several plans were located in religiously conservative areas, which they felt that their ability is strongly inhibited to influence providers. Providers may belong to the same church as a patient and erroneously not believe that the patient is at risk. Several of the plans commented that they were well aware that their state’s sexually transmitted disease rates belied the cultural beliefs, but that it was difficult to convince providers—particularly older providers—that screening was generally appropriate.
Ct Screening Versus Cervical Cancer Screening
Only 2 plans reported that Ct and cervical cancer screenings were treated the same way. One of these plans, however, reported that this meant that there were no efforts for either. The other plan targeted both equally. One plan reported that greater efforts were made for Ct screening because the cervical cancer screening rate was 84%. All other plans reported treating the 2 quite differently and favoring cervical cancer screening in their efforts.
One plan included cervical cancer in its P4P plan, but not Ct screening (although the plan reported considering the 2 equally important and that efforts to ensure that screening for both were similar). The other plans reported that targeting cervical cancer was easier because of the perception of a link between sexual activity and Ct screening but not between sexual activity and cervical cancer. Samples of reasons included the following:
* “It’s easier to discuss cervical cancer with members.”
* Providers do not “push back” on cervical cancer screening but do for Ct screening.
* “It’s harder to target Ct screening due to its link to sexual activity. Cervical cancer is easier to target as it is not as closely linked to sexual activity and is a cancer.”
* “People are willing to talk about cervical cancer, but not Ct.”
* “Ct screening is a provider based issue and cervical cancer is a member based issue.”
* “Unlike Ct screening, no one has an issue with addressing cervical cancer screening.”
One plan reported packaging Ct screening and breast and cervical cancer screening as a unit. Most plans, however, reported much more extensive efforts associated with cervical cancer screening. In addition to the P4P plan mentioned previously, another plan calculated and published cervical cancer screening rates and provided financial rewards to high-performing plans. Plans reported using voice response calls to members who had not had cervical cancer screenings, profiling members, and contacting them by telephone to suggest they go in for cervical cancer screening. Overall, it was clear that the plans were much more willing to aggressively approach members to encourage cervical cancer screening.
The purposes of this study are (a) to ascertain what activities commercial health plans have attempted (if any) to increase their Ct screening rates and (b) to identify barriers to improving Ct screening rates in commercial health plans. In terms of activities, we found that most plans were aware of Ct screening and had activities in place designed to improve their Ct screening rates. All plans had practice guidelines in place, generally based on national standards. These standards also mirror the HEDIS standards, so the plans were recommending in their practice guidelines that providers meet the HEDIS standard. Plans also had implemented a wide range of interventions, ranging from generic newsletters or fact sheets outlining Ct screening recommendations to in-depth analysis of provider’s performance on Ct screening rates versus other providers to letters to providers identifying women by name who needed to receive Ct screening. Despite this, the median screening rates among the plans were only 35%.
Many of the efforts described by the low-performing plans mirror activities reported by high-performing plans. The key question is why the high-performing plans were successful and the low-performing plans were not. There are many possibilities for the differences. One is that although the high-performing plans may have engaged in similar activities, the high-performing plans were simply more effective at implementation. Another is that although the activities are ostensibly similar, the high-performing plan activities were in practice more extensive.
The plans themselves identified numerous barriers to improving their Ct screening rates. However, it is important to note that most of these limitations affect all health plans, not just low-performing plans. If the HEDIS Ct screening measure is flawed, for example, overestimating the size of the denominator, then the screening measure should be low for all health plans. Similarly, if women seek care from public health departments, this would be expected to affect all health plans. The barriers described by low-performing plans generally do not explain why the interviewed plans do badly relative to other plans.
One barrier cited by low-performing plans that was echoed by the high-performing plans was the difficulty in reaching out to minors. Given that many parents are unaware and/or uncomfortable with minors being sexually active and that the same parents are typically the primary insured member, plans in both surveys reported very limited outreach to girl younger than 18.
We consistently found that health plans are reluctant to approach minors to encourage Ct screening tests. Although this may serve as a justification for low-performing plans, it may also create bias in the HEDIS measures in terms of measuring plan quality. If there is a segment of the population with whom plans are unable (or unwilling) to intervene, then the measure will, to some extent, simply reflect underlying practice patterns rather than act as a measure of the health plan’s quality. Although all measures reflect, to some degree, underlying practice patterns, if the plan cannot (or will not) change the practice pattern, then the measure will provide no information about the relative merits of one plans versus another. This will also serve to minimize the criticism that the measure is flawed because of the number of adolescents who are not sexually active but are prescribed birth control pills.
Confidentiality was a key concern reported by health plans. Separate concerns such as testing minors and women seeking testing at a public health clinic revolve around the issue of ensuring that the screening test is confidential. A number of high-performing plans reported changing their coding so that a Ct screening test was reported as “Laboratory Services.” This is a low-cost change that could potentially encourage Ct screening among some women. Work on this topic is also being actively carried out by the National Chlamydia Coalition, which has formed a Special Policy Group on Adolescent Confidentiality and developed tools for providers to use to obtain minor’s consent and ensure confidentiality.
Plans expressed skepticism about the accuracy of the Ct screening measure. Although we do not believe that flaws in the measure can reasonably explain differences among health plans because the flaws would apply equally to low- and high-performing plans, studies validating the measure would be useful to convince low-performing plans that the low rates reflect in HEDIS measure reflect low rates of actual screening.
Cervical cancer has much higher screening rates and less social stigma, although cervical cancer is also associated with sexual activity. If Ct screening can be incorporated into a standard “women’s health” screening protocol, there may be a reduction in stigma for Ct screening that would increase screening rates.
Finally, a number of plans reported that they believed the plans were helpless to increase Ct screening because the individual plans were too small to change provider behavior. The informants believed that coordinated efforts by all the health plans in an area could change provider behavior, but coordination among plans on an issue such as Ct screening is extremely difficult. This suggests a role for state public health departments and the CDC to work with health plans to coordinate efforts in this area. However, other plans reported that, with the correct messaging, plans could convince providers to change their behavior. The most successful messaging talked about the link between CT and infertility. When the medical director emphasized that link to providers, he reported “that’s when the light begins to go on that (doctors) should make it a part of their routine.”