Date of Award

January 2013

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Robert Tigelaar

Second Advisor

Toby Maurer

Subject Area(s)




Authors: Amy Forrestel, Erin Amerson, Lisa Butler, Jeffrey Martin, Toby Maurer. Dept of Dermatology, University of California San Francisco, San Francisco, CA. (Sponsored by Robert Tigelaar, MD. Dept of Dermatology, Yale University School of Medicine, New Haven, CT)

In East Africa, HIV-associated Kaposi Sarcoma (KS) has high prevalence, morbidity, and mortality. Despite visible skin lesions, patients with KS often do not obtain a formal diagnosis until they have developed advanced disease. This thesis explores three hypothetical barriers to and causes of delays to diagnosis of KS in Uganda in Kenya: patient- and healthcare-system driven delays in healthcare seeking and diagnosis, lack of routine practice of skin biopsy to confirm clinically suspected KS, and inaccurate histopathologic diagnosis by local pathologists. Accordingly, the specific aims of this thesis are: 1. Perform qualitative interviews with patients with advanced disease to identify barriers to obtaining care and diagnosis; 2a. Examine the accuracy of clinical diagnosis of KS; 2b. Evaluate the accuracy of histopathologic diagnosis made by local African pathologists. For aim 1, semi-structured interviews were performed on 20 patients, and descriptive thematic coding was the primary analytic strategy. For aims 2a and 2b, biopsies were performed on patients with clinical suspicion of KS. For 2a, this clinical diagnosis was compared to histopathologic diagnosis by dermatopathologists at UCSF (N = 739). For 2b, local African pathologist diagnoses were compared to dermatopathologist diagnoses at UCSF (N=1074). Results from aim 1 showed the median time from first noticing lesions to receiving a KS diagnosis at a facility capable of administering antiretroviral therapy was 7 months. Determinants of delayed diagnosis fell into 3 categories: patient-related, logistical, and system-related. The most common issues were lack of concern over symptoms; out-of-pocket costs to access care, initial use of traditional healers; difficulties in achieving contact with medical providers; lack of coordination in the referral process; and provider misdiagnosis or miscommunication.

Results from aim 2 revealed an overall accuracy of 71% for clinical diagnosis of KS, and a concordance between African pathologists and US dermatopathologists of 71%with a sensitivity of 72% and specificity of 84%, when the US dermatopathologist was considered the gold standard. Other diagnoses identified by biopsy reports when the specimens were not KS included: scar, wart, post-inflammatory pigment change, psoriasis, lymphoma, bacillary angiomatosis, morphea, sarcoidosis, polyarteritis nodosa, pyogenic granuloma, mycobacterial dermatitis, lichen planus, secondary syphilis, erythema multiforme, eccrine poroma, and xanthoma. A range of factors contributing to delays in diagnosis of KS in Uganda and Kenya were uncovered in this thesis. These ranged from subjective factors identified by patients about their experience in health-seeking, and objective data on the accuracy of clinical and histopathologic diagnosis. It appears that an intervention strategy to promote early diagnosis will need to target this entire spectrum of barriers and include patient and physician education, and increased training of pathologists in Africa.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.