Printer-Friendly Version (104k, Adobe Acrobat)

Back    More

Rationale for Screening

In the last 50 years medical screening programs have been developed to address a variety of clinical and public health problems. The purpose of screening is to provide a low-cost, accessible means of determining who in a population is likely to have a certain disease and who is not.

Criteria for deciding whether or not screening for a particular disease is appropriate include:

• Is the disease an important public health problem?
• Is the natural history of the disease understood?
• Is there a recognizable latent or early symptomatic stage?
• Is there an accepted treatment for early disease?
• Is there a consensus on whom to treat?
• Are facilities for diagnosis and treatment readily available and accessible?
• Would screening be sustainable?

While cervical cancer clearly qualifies as an important public health problem in developing countries, in many of these countries effective low-cost screening programs are not readily available.

To date, most cervical cancer prevention efforts worldwide have focused on screening at-risk women using Papanicolaou (Pap) smears and treating precancerous lesions. Where there is quality screening and coverage of the population is high (70% or more), these efforts have helped reduce invasive cervical cancer incidence by as much as 90%. When coverage is high, it is not necessary to screen women frequently to have an impact on disease incidence. For example, the approach of screening all women once every 5 years, and treating all those with dysplasia, would reduce the estimated incidence of cervical cancer by about 84% (Table 1). Even screening every 10 years would reduce the incidence by an estimated 64%. Where there are technical, logistic and other barriers to effective provision of Pap screening and followup care, however, screening programs have had little effect on disease incidence (Sankaranarayanan and Pisani 1997).

Table 1. Reduction in Cumulative Cervical Cancer Rate with Different Frequencies of Screening 

Frequency of Screeninga	% Reduction in Cumulative Rate
1 year 2 years 3 years 5 years 10 years	93.5 92.5 90.8 83.6 64.1

^aScreening all women age 35–64 who have had at least one previous negative Pap smear.

Source: IARC 1986. 

Preventing infection with HPV also will prevent cervical cancer. This primary prevention approach, however, presents greater challenges than for most other sexually transmitted diseases (STDs). While the standard STD prevention recommendations (such as regular use of condoms or other barrier methods) will help some women avoid HPV infection, the degree to which they will affect the overall incidence of cervical cancer is unclear (see Chapter 2 for additional details).

Limitations of Cervical Cancer Screening in Developing Countries 

If the burden of disease is high and screening is known to be a cost-effective intervention, why is screening not more widely available in developing countries? One reason relates to the nature of the Pap smear, the most common screening test. Cytology-based screening programs using Pap smears are complex and costly. Performing a Pap test may seem relatively simple, but numerous steps are required to take an adequate smear, process and analyze the specimen and inform women of the results. If any of these steps are unreliable or logistically burdensome, the entire screening program breaks down and with it the potential for any public health benefit (Gaffikin et al 1997).

Unfortunately many, if not all, of these steps can be problematic in low-resource settings. For example, in many countries Pap smears are usually offered only in urban areas by a small private sector or at referral facilities. And, even in these settings, trained cytotechnicians and cytopathologists are scarce, while turnaround times for processing and reading specimens are long. Thus, women do not receive their results promptly and followup losses are high. Given this reality, if screening and subsequent treatment is to have a measurable effect on decreasing the incidence of disease, it needs to be based on an approach other than just Pap smears.

In the next few years, new approaches to screening women at risk of cervical cancer will likely become more widely available. These include visual screening to identify cervical lesions without reliance on Pap smears, HPV tests to identify women at high risk for cervical cancer and automated cytology screening machines to identify subsets of Pap smears that should be examined by cytologists.¹ Of these, recent data on visual inspection of the cervix using acetic acid (VIA) suggest this approach may be at least as effective as Pap smears in detecting disease and is associated with fewer logistic and technical constraints.

History of Research on Visual Inspection 

Before the advent of Pap smears and large-scale screening programs, physicians relied on looking at the cervix to detect abnormalities. For example, the Schiller test (i.e., application of dilute aqueous iodine solution to the cervix to aid in differentiating "mature" normal from "immature" abnormal cells) was used for many years. After the 1950s, when the Pap smear became the standard for cervical cancer screening, the increasing numbers of women undergoing this test led to increased utilization of the colposcope to confirm abnormal tests. Because of the expense and inconvenience of colposcopy services, clinicians began to explore whether unmagnified visualization of the cervix could be used as an adjunct to cytology so that only women truly in need of colposcopy could be identified more effectively and efficiently.

In 1982, Ottaviano and La Torre published a revealing study involving 2,400 women who were examined visually and colposcopically after a cervical wash with acetic acid.² A key result was that "naked-eye" (unmagnified) inspection detected abnormalities in 98.4% of the cases (i.e., 307 of 312 women assessed colposcopically as having an abnormal squamocolumnar junction). In addition, using VIA, 98.9% of cases were identified as normal (i.e., in 1,568 of 1,584 women diagnosed as normal by colposcopy). These authors concluded that "colposcopic magnification is not essential in clinical practice for the identification of the cervix ‘at risk.’"

In two more recent studies, it was demonstrated that VIA can be helpful in reducing referrals for colposcopy without compromising quality of care. For example, Slawson et al (1992) found that among women who eventually had an abnormal biopsy, VIA detected disease in approximately 64% of such cases, a rate comparable to what they found for the Pap smear (68%). In addition, as the investigators became more experienced, the positive predictive value improved by almost 30%. Thus the authors concluded that "[VIA] is a safe, simple and effective adjunct to the Papanicolaou smear for cervical cancer screening." Likewise, Frisch et al (1994) discovered that using Pap smears together with VIA yielded better results (i.e., if the Pap smear and VIA both were normal, colposcopy and biopsy were also more likely to be normal) than if just a Pap smear was performed.

Visual Inspection as an Alternative to Pap Smears in Low-Resource Settings 

Even though these studies suggested the potential value of VIA as a screening approach, evidence from rigorous scientific studies in developing country settings was needed before clinicians would likely accept VIA as an alternative to cytology—even in settings where Pap smear-based services were not possible. To this end, in 1994 a study was conducted in South Africa involving VIA and Pap smears performed in a mobile unit equipped to process smears on site (Megevand et al 1996). In this study, a gynecologist performed colposcopy to confirm disease either immediately or within a few days after screening. The positive predictive value for VIA was found to be similar to that of Pap smears and the authors concluded that "naked-eye visualization of the cervix after application of diluted acetic acid...warrants consideration as an alternative to cytologic screening." In 1997, Sankar et al described an experience in India with women who agreed to have VIA, cervicoscopy and cytology. Preliminary results of this study indicated that, compared to cytology, VIA was more sensitive (91.2% versus 75% for Pap smears) in detecting severe dysplasia or worse lesions (Sankaranarayanan et al 1997).

The question of whether or not VIA can effectively distinguish diseased from nondiseased cervixes when compared to Pap smears under the same field conditions was also addressed in a large-scale study (more than 10,000 women) in Zimbabwe that compared VIA to Pap smears under similar circumstances. In Phase 2 of this study, where direct test quality estimates were calculated, the reported sensitivity (77%) of VIA was higher than that of the Pap smear whereas specificity (64%) was lower (University of Zimbabwe/JHPIEGO Cervical Cancer Project 1999) (Table 2).³

Table 2. Test Qualities of VIA in Primary Healthcare Setting (Phase 2) 

Test	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
VIA (n=2,130)	77 (70-82)	64 (62-66)	19	96
Pap Smear (n=2,092)	44 (35-51)	91 (37-51)	33	94
PPV = Positive predictive value   NPV = Negative predictive value

A major finding from the Zimbabwe study was that nonphysicians (nurse-midwives) quickly learned to perform VIA in a primary healthcare setting and correctly identify women with no disease, those suitable for immediate treatment and those requiring referral for advanced disease. The key to their performance was their training. The week-long competency-based training involved use of a specially designed VIA cervical atlas and repeated practice on pelvic models prior to working with patients. The nurse-midwives also received supplemental (two-day) training in the work setting during the first few months of the project.

At least three other rigorous studies have been completed to date which confirm VIA’s usefulness as a screening tool in low-resource settings (Table 3). Based on the results of the Zimbabwe study and those of the more recent studies in South Africa, China and India, VIA represents a proven, simple alternative means of identifying women with precancerous cervical lesions (Kitchener and Symonds 1999, Parkin and Sankaranarayanan 1999).

In summary, VIA can be used in low-resource settings because:

• it can effectively identify most precancerous lesions,
• it is noninvasive, easy to perform and inexpensive,
• it can be performed by all levels of healthcare workers in almost any setting,
• it provides immediate results on which decisions regarding treatment or referral can be made, and
• all required supplies and equipment are readily available locally.

Table 3. Test Qualities of VIA When Performed as Primary Screening Method in Low-Resource Settings

 

Study	Country	Number of Cases	Detection of HGSIL and Cancer
			Sensitivitya	Specificitya
Megevand et al (1996)	South Africa	2,426	65%	98%
Sankaranarayanan et al (1998)	India	2,935	90%	92%
Sankaranarayanan et al (1999)	India	1,351	96%	68%
University of Zimbabwe/JHPIEGO (1999)	Zimbabwe	2,148	77%	64%
Belinson (unpublished)	China	1,997	71%	74%
Denny et al (unpublished)	South Africa	2,944	67%	84%
Sankaranarayanan and Wesley (unpublished)	India	2,462	84%	90%

^a Estimated from the number provided in the manuscript and does not reflect adjustment(s) for verification bias.

Adapted from: Megevand et al (1996), Sankaranarayanan et al (1998), Sankaranarayanan et al (1999), University of Zimbabwe/JHPIEGO Cervical Cancer Project (1999), Belinson (unpublished), Denny et al (unpublished) and Sankaranarayanan and Wesley (unpublished).

---

¹ Appendix A lists a number of cervical cancer screening tests, their technical components, benefits and limitations.

² Acetic acid is absorbed by immature cells, turning their cytoplasm cloudy. To the human eye, this reaction looks white and is referred to as an "acetowhite" change; the tissue itself is often referred to as white epithelium. Mature squamous cells and glandular cells do not react this way.

³Appendix B describes the commonly measured screening test qualities and factors affecting their interpretation.

Back    More