Report on basic findings in outpatient facility evaluations in Arusha Municipality, Arumeru District and Monduli District, Arusha Region, Tanzania : Report on basic findings in outpatient facility evaluations in Arusha Municipality, Arumeru District and Monduli District, Arusha Region, Tanzania

Report on Basic Findings in Outpatient Facility Evaluations in Arusha Municipality, Arumeru District and Monduli District, Arusha Region, Tanzania Kenneth L. Leonard Melkiory Masatu June 10, 2003 Department of Economics MC 3308, 420 W 118th Street, Columbia University, New York, New York 10027 KL206@columbia.edu CEDHA, Arusha TZ cmasatu@cedha.ac.tz  Contents Front Page 1 1 Introduction 6 2 Absenteeism among medical personnel 8 2.1 Comparing postings to standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2 Scheduled personnel: Present and not present . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3 Physical Infrastructure and Equipment Availability 21 4 Pharmaceutical Availability 24 5 Examination Room Evaluation and Clinician Qualifications 27 6 Nursing Quality (Dispensing, Injections and Dressing) 29 7 Consultation Quality 31 8 Consultation Quality (Using Vignettes) 45 8.1 The use of vignettes as a quality evaluation technique . . . . . . . . . . . . . . . . . . . . . . 45 Explanation of scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 8.2 Results using Evaluation with Vignettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 8.3 Vignette Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 8.3.1 Labratory Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 8.3.2 Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 8.3.3 Treatment and Prescription Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 9 Exit Interview: Visit Reasons and Quality Evaluation 59 9.1 Patient opinion compared to research team evaluation . . . . . . . . . . . . . . . . . . . . . . 59 9.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Glossary 64 A Instruments 77 A.1 Facility Evaluation Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 A.2 Vignettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 A.2.1 Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 A.2.2 Sample Vignette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 A.2.3 Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 88 Potential Physical Examination Questions . . . . . . . . . . . . . . . . . . . . . . . . . 89 A.2.4 Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 89 Potential Physical Examination Questions . . . . . . . . . . . . . . . . . . . . . . . . . 91 A.2.5 Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 91 Physical Examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 A.2.6 Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 92 Physical Examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 A.2.7 Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 93 Physical Examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 A.2.8 Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Potential History Taking Questions and their Response . . . . . . . . . . . . . . . . . 94 2  Physical Examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 A.3 Vignette Score Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 A.3.1 Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 A.3.2 Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 A.3.3 Druguse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 A.3.4 Laboratory Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 List of Tables 1 Facitlities Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2 Number of personnel posted to facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3 Posted personnel compared to government standards by district level and system . . . . . . . 10 4 Posted personnel compared to government standards by zone level and system . . . . . . . . 11 5 Regression analysis of posted personnel compared to government standards . . . . . . . . . . 12 6 Scheduled personnel present and not present by district level and system: Nurses . . . . . . . 14 7 Scheduled personnel present and not present by district level and system: Clinicians . . . . . 15 8 Scheduled personnel present and not present by district level and system: Clinicians and Nurses 16 9 Scheduled personnel present and not present by zone level and system: Nurses . . . . . . . . 17 10 Scheduled personnel present and not present by zone level and system: Clinicians . . . . . . . 18 11 Scheduled personnel present and not present by zone level and system: Clinicians and Nurses 19 12 Regression analysis of present and not present personnel . . . . . . . . . . . . . . . . . . . . . 20 13 Facility Infrastructure by District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 14 Facility Infrastructure by Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 15 Facility Infrastructure by Owner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 16 Pharmaceutical Availability Evaluation by District . . . . . . . . . . . . . . . . . . . . . . . . 24 17 Pharmaceutical Availability Evaluation by Level . . . . . . . . . . . . . . . . . . . . . . . . . 25 18 Pharmaceutical Availability Evaluation by Owner . . . . . . . . . . . . . . . . . . . . . . . . . 25 19 Pharmaceutical Availability by Quartile of Last Delivery of Drugs . . . . . . . . . . . . . . . 26 20 Clinician Qualification and Examining Room Characteristics by District . . . . . . . . . . . . 27 21 Clinician Qualification and Examining Room Characteristics by Level . . . . . . . . . . . . . 28 22 Clinician Qualification and Examining Room Characteristics by Owner . . . . . . . . . . . . 28 23 Nursing Quality (Dispensing, Injections and Dressing) by District . . . . . . . . . . . . . . . . 29 24 Nursing Quality (Dispensing, Injections and Dressing) by Facility Level . . . . . . . . . . . . 30 25 Nursing Quality (Dispensing, Injections and Dressing) by Facility Owner . . . . . . . . . . . . 30 26 Consultation Quality by District: Details (part I) . . . . . . . . . . . . . . . . . . . . . . . . . 32 27 Consultation Quality by District: Details (part II) . . . . . . . . . . . . . . . . . . . . . . . . 33 28 Consultation Quality by District: Total Scores . . . . . . . . . . . . . . . . . . . . . . . . . . 34 29 Consultation Quality by Facility Level: Details (part I) . . . . . . . . . . . . . . . . . . . . . 34 30 Consultation Quality by Facility Level: Details (part II) . . . . . . . . . . . . . . . . . . . . . 35 31 Consultation Quality by Facility Level: Total Scores . . . . . . . . . . . . . . . . . . . . . . . 36 32 Consultation Quality by Facility Owner: Details (part I) . . . . . . . . . . . . . . . . . . . . . 36 33 Consultation Quality by Facility Owner: Details (part II) . . . . . . . . . . . . . . . . . . . . 37 34 Consultation Quality by Facility Owner: Total Scores . . . . . . . . . . . . . . . . . . . . . . 38 35 Consultation Quality by Cadre: Details (part I) . . . . . . . . . . . . . . . . . . . . . . . . . . 38 36 Consultation Quality by Cadre: Details (part II) . . . . . . . . . . . . . . . . . . . . . . . . . 39 37 Consultation Quality by Cadre: Total Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 38 Consultation Quality by Quartile of Experience: Details (part I) . . . . . . . . . . . . . . . . 40 39 Consultation Quality by Quartile of Experience: Details (part II) . . . . . . . . . . . . . . . . 41 40 Consultation Quality by Quartile of Experience: Total Scores . . . . . . . . . . . . . . . . . . 42 41 Consultation Quality by Quartile of Experience and Cadre: Details (part I) . . . . . . . . . . 42 42 Consultation Quality by Quartile of Experience and Cadre: Details (part II) . . . . . . . . . . 43 43 Consultation Quality by Quartile of Experience and Cadre: Total Scores . . . . . . . . . . . . 44 44 Vignette Quality by District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 45 Vignette Quality by Facility Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 46 Vignette Quality by Facility Owner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3  47 Vignette Quality by Cadre . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 48 Vignette Quality by Quartile of Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 49 Number of Labtests ordered for each Vignette . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 50 Labtest use by Owner Level breakdown: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . 51 51 Labtest use by Owner Level breakdown: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . 51 52 Labtest use by Owner Level breakdown: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . 52 53 Labtest use by Owner Level breakdown: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . 52 54 Labtest use by Owner Level breakdown: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . 52 55 Labtest use by Owner Level breakdown: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . 52 56 Diagnosis by Owner Level breakdown: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . 52 57 Diagnosis by Owner Level breakdown: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . 53 58 Diagnosis by Owner Level breakdown: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . 53 59 Diagnosis by Owner Level breakdown: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . 53 60 Diagnosis by Owner Level breakdown: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . 53 61 Diagnosis by Owner Level breakdown: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . 54 62 Drugs (treatment) used by Owner Level breakdown: Vignette #1 . . . . . . . . . . . . . . . . 55 63 Drugs (treatment) used by Owner Level breakdown: Vignette #2 . . . . . . . . . . . . . . . . 56 64 Drugs (treatment) used by Owner Level breakdown: Vignette #3 . . . . . . . . . . . . . . . . 56 65 Drugs (treatment) used by Owner Level breakdown: Vignette #4 . . . . . . . . . . . . . . . . 57 66 Drugs (treatment) used by Owner Level breakdown: Vignette #5 . . . . . . . . . . . . . . . . 57 67 Drugs (treatment) used by Owner Level breakdown: Vignette #6 . . . . . . . . . . . . . . . . 57 68 Number of Drugs prescribed for each Vignette . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 69 Visit Reason and Quality Evaluation by District . . . . . . . . . . . . . . . . . . . . . . . . . 59 70 Visit Reason and Quality Evaluation by Facility Level . . . . . . . . . . . . . . . . . . . . . . 60 71 Visit Reason and Quality Evaluation by Facility Owner . . . . . . . . . . . . . . . . . . . . . 60 72 Patient opinion on drug availability and Drug Evaluation . . . . . . . . . . . . . . . . . . . . 61 73 Patient Opinion of Nursing Quality compared to Evaluation of Nursing Quality . . . . . . . . 62 74 Patient Opinion of Consultation Quality and Research Team Evaluation . . . . . . . . . . . . 62 75 Patient Opinion of Consultation Quality (Facility Average) and Research Team Evaluation (Facility Average) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 76 Patient Opinion of Consultation Quality and Vignette Quality . . . . . . . . . . . . . . . . . 62 77 Diagnosis and Diagnosis Score: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 78 Diagnosis and Diagnosis Score: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 79 Diagnosis and Diagnosis Score: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 80 Diagnosis and Diagnosis Score: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 81 Diagnosis and Diagnosis Score: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 82 Diagnosis and Diagnosis Score: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 83 Treatment and Treatment Score: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 84 Treatment and Treatment Score: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 85 Treatment and Treatment Score: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 86 Treatment and Treatment Score: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 87 Treatment and Treatment Score: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 88 Treatment and Treatment Score: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 89 Treatment and Druguse Score: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 90 Treatment and Druguse Score: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 91 Treatment and Druguse Score: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 92 Treatment and Druguse Score: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 93 Treatment and Druguse Score: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 94 Treatment and Druguse Score: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 95 Labtest use and Labtest Score: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 96 Labtest use and Labtest Score: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 97 Labtest use and Labtest Score: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 98 Labtest use and Labtest Score: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 99 Labtest use and Labtest Score: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 100 Labtest use and Labtest Score: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 4  List of Figures 1 Patient Permission Card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 2 Consultation Observation: Cover Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 3 Consultation Observation: Evaluation (page 1) . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4 Consultation Observation: Evaluation (page 2 & 3) . . . . . . . . . . . . . . . . . . . . . . . . 80 5 Consultation Observation: Evaluation (page 4) . . . . . . . . . . . . . . . . . . . . . . . . . . 81 6 Nursing Evaluation: Drug Dispensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 7 Nursing Evaluation: Injection Administration & Wound Dressing . . . . . . . . . . . . . . . . 83 8 Infrastructure and Equipment Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 9 Pharmacy Stock Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 10 Exit Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 11 Vignette Evaluation Sheet: Vignette #1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 12 Vignette Evaluation Sheet: Vignette #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 13 Vignette Evaluation Sheet: Vignette #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 14 Vignette Evaluation Sheet: Vignette #4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 15 Vignette Evaluation Sheet: Vignette #5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 16 Vignette Evaluation Sheet: Vignette #6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5  1 Introduction This report details the findings of follow up visits to 40 health facilities in the Arusha region conducted during the months of February, March and April of 2003. An effort was made to visit facilities on days when there were more likely to be patients attending (clinic days or markets days), though it was not possible to do this with each facility. Three of the facilities have been closed in the last few years, and at one facility Table 1: Facitlities Examined Code Facility Ownership Type District Division Note A01 Ngarenaro Govt Disp Arusha Municip A02 Levolosi Govt HCenter Arusha Municip A03 St. Elizabeth RC Hospital Arusha Municip A04 Kaloleni Govt HCenter Arusha Municip A05 Mt. Meru Hosp. Govt Hospital Arusha Municip A06 Njiro SDA Disp Arusha Municip Closed A07 Makao Mapya SDA Disp Arusha Municip Closed AM01 Selian Luth Hospital Arumeru Muklat AM02 ArumeruHosp. Govt Hospital Arumeru Poli AM03 Olkokola Govt Disp Arumeru Muklat AM04 Oldonyo Sambu Govt Disp Arumeru Muklat AM05 Octurmet Govt HCenter Arumeru Muklat AM06 Kisongo Luth Disp Arumeru Muklat AM07 Kisongo2 SDA Disp Arumeru Muklat AM08 Olkokola RC Disp AruMeru Muklat AM09 Musa Govt Disp Arumeru Muklat AM10 Mwandet Govt Disp Arumeru Muklat AM11 Ngaremtoni COGI Disp AruMeru Muklat AM12 Oldonyo Sambu COGI Disp AruMeru Muklat M01 Monduli Govt Hospital Monduli Kisongo M02 Namanga Govt Disp Monduli Longido M03 MonduliJuu Govt Disp Monduli Kisongo M04 Lepurko Govt Disp Monduli Kisongo M05 Longido Govt HCenter Monduli Longido M06 Kimokowa Luth Disp Monduli Longido M07 Engarenaibor Govt Disp Monduli Longido M08 Arkatan Govt Disp Monduli Kisongo M09 Engikariet RC Disp Monduli Longido M10 Kitumbeine Luth Disp Monduli Longido M11 Elangatadapash Govt Disp Monduli Longido M12 Gelai Lumbwa Luth Disp Monduli Longido M13 Gelai Bomba Govt Disp Monduli Longido No health worker found M14 TMA Govt Disp Monduli Kisongo M15 Mferiji Govt Disp Monduli Kisongo M16 Elwai Luth Disp Monduli Kisongo Closed M17 Mundarara Priv Phmy Monduli Longido M18 Mundarara Govt Disp Monduli Longido M19 Namanga COGI Disp Monduli Longido PT01 AICC Parastatal Hospital Arusha Municip PT02 Ithna Asheri Islamic Hospital Arusha Municip we found no one present. Otherwise, the staff were receptive and cooperative with our research team. In addition, patients, particularly in the rural areas, were happy to see facilities being evaluated. We sought permission from every patient before the quality of the services they were receiving was evaluated and in almost every single case this permission was granted. We evaluated the facility according to the standards that would be expected of a dispensary. We looked 6  for supplies and infrastructure that would be expected at the basic dispensary level, and we examined the competence of nurses and clinicians for procedures and illnesses that could be treated at a dispensary. Thus, all facilities were evaluated on the same basis, whether dispensary, health center or hospital. Thus, if we conclude that hospital is well equipped this does not mean that they are as well as equipped as they should be since we did not examine supplies or equipment that would be expected at a hospital. At each facility, we used the surveys shown as Figures 1 through 16 in appendix A.1. The first stage of the research was to obtain the permission of patients to observe their consultation, as well as drug dispensing, injections or wound dressing if necessary. To do this we read the text on the patient card (Figure 1) and then gave them this card (which also contained a patient number with which to identify patients.) Patients were understanding and few objected to our observing the procedures. The fact that all the researchers were medical personnel made patients more comfortable. Many stated that they were pleased to see people asking questions about quality. Once the patient had agreed to be observed we watched them receive services in consultation (Figures 3, 4 and 5 with results shown in Section 7) drug dispensing, injections and wound dressing (Figure 6 and Figure 7 with results shown in Section 6) and finally they were asked to respond to an exit interview (Figure 10 with results shown in Section 9). In addition each facility was evaluated for physical infrastructure (Figure 8 with results shown in Sec- tion 3) and drug availability (Figure 9 with results shown in Section 4). Each physician was evaluated for quality by the team using vignettes as shown in Figure 11 through 16 explained in Section 8.1 with results shown in Section 8.2. 7  2 Absenteeism among medical personnel The first item of quality is the presence of medical personnel at the facility. With this in mind we evaluated all facilities according to two basic criteria, were the proper medical personnel posted to the facility, and were the personnel posted actually present? Table 2: Number of personnel posted to facilities hospital health center dispensary type total avg min total avg min total avg min Number of Facilities 7 5 24 Medical Officer 12 1.71 3 Assistant Medical Officer 12 1.71 4 0.80 1 1 0.04 Clinical Officer 40 5.71 6 15 3.00 4 10 0.42 1 Clinical Assistant 4 0.57 8 1.60 13 0.54 1 All Clinicians 68 9.71 9 27 5.40 5 24 1.00 2 Nursing Officer 15 2.14 Public Health Nurse (Grade B) 10 1.43 8 1.60 2 2 0.08 1 Registered Nurse Midwife 60 8.57 10 2.00 2 6 0.25 1 MCH aide 8 1.14 19 3.80 2 13 0.54 1 All Nurses 93 13.29 37 7.40 6 21 0.88 3 Radiographer 6 0.86 Laboratory Technician 11 1.57 2 1 0.04 Laboratory Assistant 8 1.14 2 3 0.60 1 3 0.13 1 Pharmacist 2 0.29 1 Pharmaceutical Assistant 7 1.00 1 0.20 Medical Attendent 105 15.00 71 14.20 2 26 1.08 Health Officer 1 0.20 3 0.13 Health Assistant 5 0.21 Other Medical 9 1.29 3 0.60 1 0.04 Other Non Medical 24 3.43 7 1.40 7 0.29 Clinicians are listed in order of cadre and nurses broadly in order of cadre. A medical attendant is not qualified as either a nurse or a clinician and is a primary school leaver with basic medical training. Health officers and assistants are primarily public health officers who do outreach to the community and are not qualified as either a nurse nor a clinician. Other Medical includes dentists and dental assistants. Other non medical includes janitors and security officers. Table 2 shows the number of personnel posted to each facility we visited compared to the government mandated minimum number of personnel. The government mandate is a guideline and is not legally binding, but it is designed to be a minimum for effective functioning. It is not secret that facilities fall short on this measure, but we are trying to get an idea of the characteristics of facilities that fall short. There are three basic categories of personnel, clinicians, nurses and others. We will focus only on the clinicians and nurses. The minimum staffing of nurses in a hospital is a function of the number of beds in the hospital and this is difficult to determine accurately. Therefore we have not examined this category. In some cases facilities have cadre of above minimum qualifications and in some cases facilities have cadre of below minimum qualifications. To compare the absolute number of personnel we include the total for each category. Thus, while the average hospital has only 1.71 medical officers posted, they have 9.71 clinicians compared to a minimum number of 9. In this table we can see that hospitals and health centers are, on average, adequately staffed in the overall categories. However dispensaries are not adequately staffed for either clinicians or nurses. In many cases the dispensaries are staffed with medical attendants rather than clinicians or nurses. This is not an adequate substitute. In order to compare the staffing in various facilities we introduce the following categories: Level Hospital, Health Center and Dispensary (only the government operates health centers). District We surveyed three districts, Arusha, Arumeru and Monduli. These districts are broadly urban, semi-urban and rural, so we use these titles instead. 8  Zone We also use the category zone which is slightly different from District. Two facilities can be in the same district but in different zones, and a few facilities are in the same zone, but in different districts. Urban/City Within the limits of the regional capital. Major town Placed in a town that has a transportation hub, multiple markets, telephone access, etc (but not the regional capital). Close to trunk road Place within easy walking distance of inexpensive transportation to the regional capital (but does not include the major town category) Far from trunk road Not within walking distance of transportation, but within 2 to 3 hours of a trunk road by vehicle Very far from a trunk road At least a days travel from a main road. Owner There are three major systems operating in this area, government, church owned and private or other. Other includes parastatal, Islamic and COGI (Church of God in Christ) facilities. We call these private because their religious affiliation is for tax purposes only. There is no medical oversight given by the religious body. In contrast the facilities that we call church-owned are monitored by a diocesan medical office headed by a bishop. We show our analysis by district and zone separately. District is less detailed than zone but is informative because all government facilities within a district are governed by the same district medical officer. Thus the district category reflects the difficulties faced by the DMO as well and the difficulties of living in a remote posting. The zone category will overlap the districts to a certain degree but is also more informative for the church and private/other categories since government in these cases does not follow district boundaries. 2.1 Comparing postings to standards Table 3 and Table 4 show the number of personnel posted to a facility compared to the minimum level according to government standards. In these graphs we compared postings to standards only for the collection of clinicians and nurses. In other words this does not measure the degree to which posted personnel are less qualified than government standards demand, but the degree to which the number of clinicians and nurses falls below standards. We use three different measures: Difference The average difference between posted and standard. Some categories of facility have more than the minimum number and this might balance facilities that have less than the minimum. Num Below The average number by which a facility falls short of standards. Thus, over staffed facilities are assigned a zero and under staffed facilities a negative number. Percent Under Percent of facilities that are under staffed. We report these for clinicians and nurses separately. Nurse staffing for hospitals is not reported. Table 3 shows that 91% of dispensaries are under staffed with nurses and 83% of dispensaries are under staffed with clinicians. In contrast many fewer health centers and hospitals are under staffed. The rural district faces greater shortages of both nurses and clinicians, and the semi-rural faces greater shortages than the urban district. In fact the urban district has, on average, too many nurses and clinicians assigned. This is important because it suggests that reallocation might improve staffing. There is a net shortage, but there is also room for reallocation. It is interesting to note that under staffing is not a government problem. The problem faces government facilities, church facilities and private facilities. In both urban and semi-urban districts the government does a marginally better job of staffing facilities than do the church operated systems. In addition, in both the rural and semi-rural districts private facilities are better staffed. However in the urban district they are under staffed whereas the government and church operated facilities are over staffed. The urban private facilities are hospitals and the urban rural facilities are dispensaries so it is dangerous to infer too much, but the private facilities are allocating staff in a different manner than government or church facilities. Table 4 is the same basic table as Table 3, but by zone, not district. With the exception of the major town category, there is a progression towards under staffing as you get further from the city. There is over 9  Table 3: Posted personnel compared to government standards by district level and system Level District Owner # of facs Difference Num Below Percent Under nurses clin. nurses clin. nurses clin. Disp rural church 4 -2.50 -1.50 -2.50 -1.50 100 100 Disp rural government 9 -2.22 -1.33 -2.22 -1.33 100 100 Disp rural private/other 2 -3.00 0.00 -3.00 -0.50 100 50 Disp rural 15 -2.40 -1.20 -2.40 -1.27 100 93 Disp semi rural church 3 -2.00 -1.00 -2.00 -1.00 100 100 Disp semi rural government 4 -1.75 -0.50 -1.75 -0.50 75 50 Disp semi rural private/other 2 -1.00 -0.50 -1.00 -0.50 50 50 Disp semi rural 9 -1.67 -0.67 -1.67 -0.67 77 66 Disp church 7 -2.29 -1.29 -2.29 -1.29 100 100 Disp government 13 -2.08 -1.08 -2.08 -1.08 92 84 Disp private/other 4 -2.00 -0.25 -2.00 -0.50 75 50 HCenter rural government 1 -2.00 -2.00 -2.00 -2.00 100 100 HCenter semi rural government 1 -1.00 -1.00 -1.00 -1.00 100 100 HCenter urban government 3 3.33 1.67 -1.33 0.00 33 0 HCenter government 5 1.40 0.40 -1.40 -0.60 60 40 Hospital rural government 1 3.00 0.00 0 Hospital rural 1 3.00 0.00 0 Hospital semi rural church 1 -3.00 -3.00 100 Hospital semi rural government 1 11.00 0.00 0 Hospital semi rural 2 4.00 -1.50 50 Hospital urban church 1 0.00 0.00 0 Hospital urban government 1 1.00 0.00 0 Hospital urban private/other 2 -3.50 -3.50 100 Hospital urban 4 -1.50 -1.75 50 Hospital church 2 -1.50 -1.50 50 Hospital government 3 5.00 0.00 0 Hospital private/other 2 -3.50 -3.50 100 rural church 4 -2.50 -1.50 -2.50 -1.50 100 100 rural government 11 -2.20 -1.00 -2.20 -1.27 100 90 rural private/other 2 -3.00 0.00 -3.00 -0.50 100 50 semi rural church 4 -2.00 -1.50 -2.00 -1.50 100 100 semi rural government 6 -1.60 1.33 -1.60 -0.50 80 50 semi rural private/other 2 -1.00 -0.50 -1.00 -0.50 50 50 urban church 1 0.00 0.00 0 urban government 4 3.33 1.50 -1.33 0.00 33 0 urban private/other 2 -3.50 -3.50 100 Disp 24 -2.13 -1.00 -2.13 -1.04 91 83 HCenter 5 1.40 0.40 -1.40 -0.60 60 40 Hospital 7 0.71 -1.43 42 rural 17 -2.38 -1.00 -2.38 -1.24 100 88 semi rural 12 -1.60 0.08 -1.60 -0.83 80 66 urban 7 3.33 -0.14 -1.33 -1.00 33 28 church 9 -2.29 -1.33 -2.29 -1.33 100 88 government 21 -1.11 0.14 -1.89 -0.81 83 61 private/other 6 -2.00 -1.33 -2.00 -1.50 75 66 staffing of clinicians and nurses in the urban/city category and major town category and then mover and more severe under staffing as you mover further out. Again this suggests a poor allocation of resources. There is some complication in these figures due to the fact that nurses are not counted in hospitals. Table 5 is a regression analysis of the data presented in Table 4. The regression assumes a linear structure that is probably not valid, but we are limited by the number of observations. In other words, in 10  Table 4: Posted personnel compared to government standards by zone level and system Zone Level Owner # of facs Difference Num Below Percent Under nurses clin. nurses clin. nurses clin. Urban/City Disp private/other 1 0.00 -1.00 0.00 -1.00 0 100 Urban/City Disp 1 0.00 -1.00 0.00 -1.00 0 100 Urban/City HCenter government 3 3.33 1.67 -1.33 0.00 33 0 Urban/City Hospital church 2 -1.50 -1.50 50 Urban/City Hospital government 1 1.00 0.00 0 Urban/City Hospital private/other 2 -3.50 -3.50 100 Urban/City Hospital 5 -1.80 -2.00 60 Urban/City church 2 -1.50 -1.50 50 Urban/City government 4 3.33 1.50 -1.33 0.00 33 0 Urban/City private/other 3 0.00 -2.67 0.00 -2.67 0 100 Major Town HCenter government 1 -2.00 -2.00 -2.00 -2.00 100 100 Major Town Hospital government 2 7.00 0.00 0 Major Town Hospital 2 7.00 0.00 0 Major Town government 3 -2.00 4.00 -2.00 -0.67 100 33 Close to trunk Disp church 5 -2.20 -1.40 -2.20 -1.40 100 100 Close to trunk Disp government 7 -1.71 -0.86 -1.71 -0.86 85 71 Close to trunk Disp private/other 2 -2.50 0.50 -2.50 0.00 100 0 Close to trunk Disp 14 -2.00 -0.86 -2.00 -0.93 92 71 Close to trunk HCenter government 1 -1.00 -1.00 -1.00 -1.00 100 100 Close to trunk church 5 -2.20 -1.40 -2.20 -1.40 100 100 Close to trunk government 8 -1.63 -0.88 -1.63 -0.88 87 75 Close to trunk private/other 2 -2.50 0.50 -2.50 0.00 100 0 Far fr. trunk Disp church 1 -2.00 -1.00 -2.00 -1.00 100 100 Far fr. trunk Disp government 2 -2.50 -1.00 -2.50 -1.00 100 100 Far fr. trunk Disp private/other 1 -3.00 -1.00 -3.00 -1.00 100 100 Far fr. trunk Disp 4 -2.50 -1.00 -2.50 -1.00 100 100 Far fr. trunk church 1 -2.00 -1.00 -2.00 -1.00 100 100 Far fr. trunk government 2 -2.50 -1.00 -2.50 -1.00 100 100 Far fr. trunk private/other 1 -3.00 -1.00 -3.00 -1.00 100 100 V. far fr. trunk Disp church 1 -3.00 -1.00 -3.00 -1.00 100 100 V. far fr. trunk Disp government 4 -2.50 -1.50 -2.50 -1.50 100 100 V. far fr. trunk Disp 5 -2.60 -1.40 -2.60 -1.40 100 100 V. far fr. trunk church 1 -3.00 -1.00 -3.00 -1.00 100 100 V. far fr. trunk government 4 -2.50 -1.50 -2.50 -1.50 100 100 Urban/City 9 2.50 -0.56 -1.00 -1.22 25 44 Major Town 3 -2.00 4.00 -2.00 -0.67 100 33 Close to trunk 15 -1.93 -0.87 -1.93 -0.93 93 73 Far fr. trunk 4 -2.50 -1.00 -2.50 -1.00 100 100 V. far fr. trunk 5 -2.60 -1.40 -2.60 -1.40 100 100 this regression, the impact of zone is the same for each system, and the impact of system is the same for each zone. It is more likely that different systems will respond differently to zones. In Table 5, there are not many statistically significant variables. Being very far away from a major road predicts average staffing levels, degree of under staffing and the percent of facilities that are under staffed, and the sign of the variables for close to trunk and far from trunk are correct. However, there is no evidence that government or church operated facilities have any different posting habits than private facilities (with this linear specification). 11  Table 5: Regression analysis of posted personnel compared to government standards Posted/Minimum difference num below percent under Coef. Std. Err Coef. Std. Err Coef. Std. Err major town 1.55 1.39 -0.01 0.60 0.17 0.22 close to trunk -1.88 1.34 -0.72 0.58 0.49 0.21 * far fr. trunk -2.14 1.55 -1.03 0.67 0.67 0.25 * v far fr. trunk -2.79 1.59 * -1.46 0.69 * 0.71 0.25 * government 1.38 0.95 0.62 0.41 -0.11 0.15 church 0.46 1.00 0.17 0.43 0.09 0.16 hospital -0.41 1.28 -0.78 0.55 -0.02 0.20 dispensary 0.01 1.38 0.40 0.60 -0.12 0.22 clinicians 0.76 0.62 1.03 0.27 * -0.10 0.10 constant -0.80 1.21 -1.99 0.52 * 0.53 0.19 * adj R-square 17% 20% 33% * significant at the 90% level for a two sided test. 12  2.2 Scheduled personnel: Present and not present We have been looking at the number of personnel who are posted to a facility. At this point we turn to the number of personnel who are physically present when they are scheduled to be present. A posted person can be either scheduled or not scheduled, and a scheduled person can be either present or not present. We collected data on the following categories: posted Assigned to facility scheduled Scheduled to be present on the day and time of our visit not scheduled Officially off work when we visited. This category is especially important for hospitals and health facilities that staff night hours. All our visits were during normal working hours. present At the facility and working or ready to work at the time we visited. present nearby In rural dispensaries, the clinician is frequently making housecalls or at home not working but prepared to come to the clinic for any emergency. When the clinician came to the clinic shortly after we arrived or was easy to find (and ready to work), this category applied. not present but frequently present When we could not find the clinician we asked people in the village or community if the clinician was often there or not. If they said yes we counted the person as not present but frequently present. not present Not present at the facility. We cannot know whether a clinician who was present nearby would have come to the facility for any patient, or if they came just for us, but it seems reasonable to give them the benefit of the doubt and we merged the present and present nearby categories into one present category. We remain of two minds about the "not present but frequently present" category. Clearly they are not present, but they are not the same category as someone who was posted and has never come to the facility. We retain this as a separate category. Table 6, Table 7 and Table 8 examine the presence of nurses, clinicians and combined doctors and nurses by district, level and system. These number are compared only to the total number of personnel scheduled to be present, not the number posted or the government standards. Table 9, Table 10 and Table 11 present the same basic data but by zone not district. Table 6 shows that dispensaries hospitals and health centers have similar rates of absenteeism among nurses, although health centers appear to do better. The rate of absenteeism between rural, semi-rural and urban districts follows the expected pattern, with absenteeism less in urban districts and worst in rural districts. Nurses in the government system are more likely to be absent than in either the church or private systems. These trends are true at both the hospital and dispensary level. Overall 81% of nurses who were supposed to be on duty were present when we visited. In many of the facilities, all nurses were present, and in the rural government hospital only 55% of posted nurses were present. Table 7 shows the same basic data but for clinicians. The rate of presence for clinicians is significantly lower than that for nurses, 73% compared to 81%. Dispensaries have a much higher rate of absenteeism than hospitals, but health centers are the worst for clinicians. Comparing rural to semi-rural to urban districts requires taking a stand on the category of "frequently present." If you consider these people as not being different from "not present" then the appropriate column is the "present" column, and the urban and semi- rural districts are similar and better than the rural district. On the other hand if they are considered as being similar to the "present" category, then the appropriate column is the "not present" column, and the urban district is superior to the semi-rural district and vastly superior to the rural district. We suggest that "frequently present" is probably closer to "not present", than to "present" for our purposes. This is because, while randomly sampling facilities, we pick up some of the frequently present as present and some of the frequently present as not present. On average, we are measuring the average number of days that people in the category are present. When they are not present they cannot serve patients. Again, considering "frequently present" as being "not present" we see that government and church facilities are not very different from each other, and marginally worse than private facilities. 13  Table 6: Scheduled personnel present and not present by district level and system: Nurses Level District Owner scheduled Present Freq Pres Not Pres Disp rural church 2 100% (0.00) 0% (0.00) 0% (0.00) Disp rural government 7 71% (0.24) 0% (0.00) 28% (0.24) Disp semi rural church 3 66% (0.33) 0% (0.00) 33% (0.33) Disp semi rural government 5 80% (0.20) 0% (0.00) 20% (0.20) Disp semi rural private/other 4 100% (0.00) 0% (0.00) 0% (0.00) HCenter rural government 3 100% (0.00) 0% (0.00) 0% (0.00) HCenter semi rural government 4 75% (0.25) 0% (0.00) 25% (0.25) HCenter urban government 26 88% (0.11) 3% (0.04) 7% (0.07) Hospital rural government 9 55% (0.28) 0% (0.00) 44% (0.28) Hospital semi rural church 2 100% (0.00) 0% (0.00) 0% (0.00) Hospital semi rural government 26 80% (0.16) 0% (0.00) 19% (0.16) Hospital urban church 9 88% (0.11) 11% (0.11) 0% (0.00) Hospital urban private/other 11 81% (0.16) 0% (0.00) 18% (0.16) Disp rural 9 77% (0.19) 0% (0.00) 22% (0.19) Disp semi rural 12 83% (0.15) 0% (0.00) 16% (0.15) Disp church 5 80% (0.20) 0% (0.00) 20% (0.20) Disp government 12 75% (0.20) 0% (0.00) 25% (0.20) Disp private/other 4 100% (0.00) 0% (0.00) 0% (0.00) HCenter government 33 87% (0.11) 3% (0.03) 9% (0.09) Hospital rural 9 55% (0.28) 0% (0.00) 44% (0.28) Hospital semi rural 28 82% (0.15) 0% (0.00) 17% (0.15) Hospital urban 20 85% (0.13) 5% (0.05) 10% (0.09) Hospital church 11 90% (0.09) 9% (0.09) 0% (0.00) Hospital government 35 74% (0.20) 0% (0.00) 25% (0.20) Hospital private/other 11 81% (0.16) 0% (0.00) 18% (0.16) rural church 2 100% (0.00) 0% (0.00) 0% (0.00) rural government 19 68% (0.23) 0% (0.00) 31% (0.23) rural private/other 0 % () % () % () semi rural church 5 80% (0.20) 0% (0.00) 20% (0.20) semi rural government 35 80% (0.16) 0% (0.00) 20% (0.16) semi rural private/other 4 100% (0.00) 0% (0.00) 0% (0.00) urban church 9 88% (0.11) 11% (0.11) 0% (0.00) urban government 26 88% (0.11) 3% (0.04) 7% (0.07) urban private/other 11 81% (0.16) 0% (0.00) 18% (0.16) Disp 21 80% (0.16) 0% (0.00) 19% (0.16) HCenter 33 87% (0.11) 3% (0.03) 9% (0.09) Hospital 57 78% (0.17) 1% (0.02) 19% (0.16) rural 21 71% (0.21) 0% (0.00) 28% (0.21) semi rural 44 81% (0.15) 0% (0.00) 18% (0.15) urban 46 86% (0.12) 4% (0.04) 8% (0.08) church 16 87% (0.12) 6% (0.06) 6% (0.06) government 80 80% (0.16) 1% (0.01) 18% (0.15) private/other 15 86% (0.12) 0% (0.00) 13% (0.12) 111 81% (0.15) 1% (0.02) 16% (0.14) Standard errors in parentheses 14  Table 7: Scheduled personnel present and not present by district level and system: Clinicians Level District Owner scheduled Present Freq Pres Not Pres Disp rural church 2 0% (0.00) 50% (0.50) 50% (0.50) Disp rural government 6 83% (0.17) 0% (0.00) 16% (0.17) Disp rural private/other 4 50% (0.33) 0% (0.00) 50% (0.33) Disp semi rural church 3 66% (0.33) 0% (0.00) 33% (0.33) Disp semi rural government 6 83% (0.17) 0% (0.00) 16% (0.17) Disp semi rural private/other 3 66% (0.33) 0% (0.00) 33% (0.33) HCenter rural government 2 0% (0.00) 0% (0.00) 100% (0.00) HCenter semi rural government 2 100% (0.00) 0% (0.00) 0% (0.00) HCenter urban government 15 53% (0.27) 13% (0.12) 33% (0.24) Hospital rural government 8 87% (0.13) 0% (0.00) 12% (0.13) Hospital semi rural church 3 100% (0.00) 0% (0.00) 0% (0.00) Hospital semi rural government 14 64% (0.25) 0% (0.00) 35% (0.25) Hospital urban church 7 85% (0.14) 14% (0.14) 0% (0.00) Hospital urban government 6 100% (0.00) 0% (0.00) 0% (0.00) Hospital urban private/other 8 100% (0.00) 0% (0.00) 0% (0.00) Disp rural 12 58% (0.27) 8% (0.08) 33% (0.24) Disp semi rural 12 75% (0.20) 0% (0.00) 25% (0.20) Disp church 5 40% (0.30) 20% (0.20) 40% (0.30) Disp government 12 83% (0.15) 0% (0.00) 16% (0.15) Disp private/other 7 57% (0.29) 0% (0.00) 42% (0.29) HCenter government 19 52% (0.26) 10% (0.10) 36% (0.25) Hospital rural 8 87% (0.13) 0% (0.00) 12% (0.13) Hospital semi rural 17 70% (0.22) 0% (0.00) 29% (0.22) Hospital urban 21 95% (0.05) 4% (0.05) 0% (0.00) Hospital church 10 90% (0.10) 10% (0.10) 0% (0.00) Hospital government 28 78% (0.17) 0% (0.00) 21% (0.17) Hospital private/other 8 100% (0.00) 0% (0.00) 0% (0.00) rural church 2 0% (0.00) 50% (0.50) 50% (0.50) rural government 16 75% (0.20) 0% (0.00) 25% (0.20) rural private/other 4 50% (0.33) 0% (0.00) 50% (0.33) semi rural church 6 83% (0.17) 0% (0.00) 16% (0.17) semi rural government 22 72% (0.21) 0% (0.00) 27% (0.21) semi rural private/other 3 66% (0.33) 0% (0.00) 33% (0.33) urban church 7 85% (0.14) 14% (0.14) 0% (0.00) urban government 21 66% (0.23) 9% (0.09) 23% (0.19) urban private/other 8 100% (0.00) 0% (0.00) 0% (0.00) Disp 24 66% (0.23) 4% (0.04) 29% (0.22) HCenter 19 52% (0.26) 10% (0.10) 36% (0.25) Hospital 46 84% (0.13) 2% (0.02) 13% (0.12) rural 22 63% (0.24) 4% (0.05) 31% (0.23) semi rural 31 74% (0.20) 0% (0.00) 25% (0.20) urban 36 77% (0.18) 8% (0.08) 13% (0.12) church 15 73% (0.21) 13% (0.12) 13% (0.12) government 59 71% (0.21) 3% (0.03) 25% (0.19) private/other 15 80% (0.17) 0% (0.00) 20% (0.17) 89 73% (0.20) 4% (0.04) 22% (0.18) Standard errors in parentheses 15  Table 8: Scheduled personnel present and not present by district level and system: Clinicians and Nurses Level District Owner scheduled Present Freq Pres Not Pres Disp rural church 4 50% (0.33) 25% (0.25) 25% (0.25) Disp rural government 13 76% (0.19) 0% (0.00) 23% (0.19) Disp rural private/other 4 50% (0.33) 0% (0.00) 50% (0.33) Disp semi rural church 6 66% (0.27) 0% (0.00) 33% (0.27) Disp semi rural government 11 81% (0.16) 0% (0.00) 18% (0.16) Disp semi rural private/other 7 85% (0.14) 0% (0.00) 14% (0.14) HCenter rural government 5 60% (0.30) 0% (0.00) 40% (0.30) HCenter semi rural government 6 83% (0.17) 0% (0.00) 16% (0.17) HCenter urban government 41 75% (0.19) 7% (0.07) 17% (0.15) Hospital rural government 17 70% (0.22) 0% (0.00) 29% (0.22) Hospital semi rural church 5 100% (0.00) 0% (0.00) 0% (0.00) Hospital semi rural government 40 75% (0.19) 0% (0.00) 25% (0.19) Hospital urban church 16 87% (0.12) 12% (0.12) 0% (0.00) Hospital urban government 6 100% (0.00) 0% (0.00) 0% (0.00) Hospital urban private/other 19 89% (0.10) 0% (0.00) 10% (0.10) Disp rural 21 66% (0.23) 4% (0.05) 28% (0.21) Disp semi rural 24 79% (0.17) 0% (0.00) 20% (0.17) Disp church 10 60% (0.27) 10% (0.10) 30% (0.23) Disp government 24 79% (0.17) 0% (0.00) 20% (0.17) Disp private/other 11 72% (0.22) 0% (0.00) 27% (0.22) HCenter government 52 75% (0.19) 5% (0.06) 19% (0.16) Hospital rural 17 70% (0.22) 0% (0.00) 29% (0.22) Hospital semi rural 45 77% (0.18) 0% (0.00) 22% (0.18) Hospital urban 41 90% (0.09) 4% (0.05) 4% (0.05) Hospital church 21 90% (0.09) 9% (0.09) 0% (0.00) Hospital government 63 76% (0.18) 0% (0.00) 23% (0.18) Hospital private/other 19 89% (0.10) 0% (0.00) 10% (0.10) rural church 4 50% (0.33) 25% (0.25) 25% (0.25) rural government 35 71% (0.21) 0% (0.00) 28% (0.21) rural private/other 4 50% (0.33) 0% (0.00) 50% (0.33) semi rural church 11 81% (0.16) 0% (0.00) 18% (0.16) semi rural government 57 77% (0.18) 0% (0.00) 22% (0.18) semi rural private/other 7 85% (0.14) 0% (0.00) 14% (0.14) urban church 16 87% (0.12) 12% (0.12) 0% (0.00) urban government 47 78% (0.17) 6% (0.06) 14% (0.13) urban private/other 19 89% (0.10) 0% (0.00) 10% (0.10) Disp 45 73% (0.20) 2% (0.02) 24% (0.19) HCenter 52 75% (0.19) 5% (0.06) 19% (0.16) Hospital 103 81% (0.15) 1% (0.02) 16% (0.14) rural 43 67% (0.22) 2% (0.02) 30% (0.22) semi rural 75 78% (0.17) 0% (0.00) 21% (0.17) urban 82 82% (0.14) 6% (0.06) 10% (0.10) church 31 80% (0.16) 9% (0.09) 9% (0.09) government 139 76% (0.18) 2% (0.02) 21% (0.17) private/other 30 83% (0.14) 0% (0.00) 16% (0.14) 200 78% (0.17) 3% (0.03) 19% (0.15) Standard errors in parentheses 16  Table 9: Scheduled personnel present and not present by zone level and system: Nurses Zone Level Owner scheduled Present Freq Pres Not Pres Urban/City Disp private/other 3 100% (0.00) 0% (0.00) 0% (0.00) Urban/City HCenter government 26 88% (0.11) 3% (0.04) 7% (0.07) Urban/City Hospital church 11 90% (0.09) 9% (0.09) 0% (0.00) Urban/City Hospital private/other 11 81% (0.16) 0% (0.00) 18% (0.16) Major Town HCenter government 3 100% (0.00) 0% (0.00) 0% (0.00) Major Town Hospital government 35 74% (0.20) 0% (0.00) 25% (0.20) Close to trunk Disp church 4 75% (0.25) 0% (0.00) 25% (0.25) Close to trunk Disp government 9 88% (0.11) 0% (0.00) 11% (0.11) Close to trunk Disp private/other 1 100% () 0% () 0% () Close to trunk HCenter government 4 75% (0.25) 0% (0.00) 25% (0.25) Far from trunk Disp church 1 100% () 0% () 0% () Far from trunk Disp government 1 0% () 0% () 100% () Very far from trunk Disp government 2 50% (0.50) 0% (0.00) 50% (0.50) Urban/City Disp 3 100% (0.00) 0% (0.00) 0% (0.00) Urban/City Hospital 22 86% (0.12) 4% (0.05) 9% (0.09) Urban/City church 11 90% (0.09) 9% (0.09) 0% (0.00) Urban/City government 26 88% (0.11) 3% (0.04) 7% (0.07) Urban/City private/other 14 85% (0.13) 0% (0.00) 14% (0.13) Major Town Hospital 35 74% (0.20) 0% (0.00) 25% (0.20) Major Town government 38 76% (0.19) 0% (0.00) 23% (0.19) Close to trunk Disp 14 85% (0.13) 0% (0.00) 14% (0.13) Close to trunk church 4 75% (0.25) 0% (0.00) 25% (0.25) Close to trunk government 13 84% (0.14) 0% (0.00) 15% (0.14) Close to trunk private/other 1 100% () 0% () 0% () Far from trunk Disp 2 50% (0.50) 0% (0.00) 50% (0.50) Far from trunk church 1 100% () 0% () 0% () Far from trunk government 1 0% () 0% () 100% () Very far from trunk Disp 2 50% (0.50) 0% (0.00) 50% (0.50) Very far from trunk government 2 50% (0.50) 0% (0.00) 50% (0.50) Urban/City 51 88% (0.11) 3% (0.04) 7% (0.07) Major Town 38 76% (0.19) 0% (0.00) 23% (0.19) Close to trunk 18 83% (0.15) 0% (0.00) 16% (0.15) Far from trunk 2 50% (0.50) 0% (0.00) 50% (0.50) Very far from trunk 2 50% (0.50) 0% (0.00) 50% (0.50) Standard errors in parentheses Table 8 shows the combined numbers for nurses and clinicians. With the combined categories health centers are not very different from dispensaries and both are worse than hospitals. Rural is worse than semi-rural and they are both worse than urban. And the government has the highest rate of absenteeism, with the private category doing the best. Table 9 shows the same results as above, but by zone rather than district. Absenteism increases as we move outwards in zones, but not smoothly. In particular nurses are less likely to be present if the facility is in a major town than if the facility is close to a major road. This might indicate that personnel in towns are more likely to report to their post, but less likely to work according to the schedule. Those in more remote areas are more likely to come to work every day once they have decided to report to the post. In the very remote areas the decision not to report to the post becomes a strong impact. The same basic pattern is observed for clinicians (Table 10) but where as being "far from a trunk road" and "very far from a trunk road" were similar for nurses, they are different for clinicians. Notice as well a pattern we will test statistically. In the urban areas, the government has a much higher rate of absenteeism than other systems, whereas in the rural areas the rate is better than other systems. Table 11 combines both nurses and clinicians. The same basic patterns appear in this graph. The more remote the greater the rate of absenteeism and the government is better in rural than it is urban areas. 17  Table 10: Scheduled personnel present and not present by zone level and system: Clinicians Zone Level Owner scheduled Present Freq Pres Not Pres Urban/City Disp private/other 1 100% () 0% () 0% () Urban/City HCenter government 15 53% (0.27) 13% (0.12) 33% (0.24) Urban/City Hospital church 10 90% (0.10) 10% (0.10) 0% (0.00) Urban/City Hospital government 6 100% (0.00) 0% (0.00) 0% (0.00) Urban/City Hospital private/other 8 100% (0.00) 0% (0.00) 0% (0.00) Major Town HCenter government 2 0% (0.00) 0% (0.00) 100% (0.00) Major Town Hospital government 22 72% (0.21) 0% (0.00) 27% (0.21) Close to trunk Disp church 3 66% (0.33) 0% (0.00) 33% (0.33) Close to trunk Disp government 8 87% (0.13) 0% (0.00) 12% (0.13) Close to trunk Disp private/other 5 40% (0.30) 0% (0.00) 60% (0.30) Close to trunk HCenter government 2 100% (0.00) 0% (0.00) 0% (0.00) Far from trunk Disp church 1 0% () 100% () 0% () Far from trunk Disp government 2 100% (0.00) 0% (0.00) 0% (0.00) Far from trunk Disp private/other 1 100% () 0% () 0% () Very far from trunk Disp church 1 0% () 0% () 100% () Very far from trunk Disp government 2 50% (0.50) 0% (0.00) 50% (0.50) Urban/City Disp 1 100% () 0% () 0% () Urban/City Hospital 24 95% (0.04) 4% (0.04) 0% (0.00) Urban/City church 10 90% (0.10) 10% (0.10) 0% (0.00) Urban/City government 21 66% (0.23) 9% (0.09) 23% (0.19) Urban/City private/other 9 100% (0.00) 0% (0.00) 0% (0.00) Major Town Hospital 22 72% (0.21) 0% (0.00) 27% (0.21) Major Town government 24 66% (0.23) 0% (0.00) 33% (0.23) Close to trunk Disp 16 68% (0.23) 0% (0.00) 31% (0.23) Close to trunk church 3 66% (0.33) 0% (0.00) 33% (0.33) Close to trunk government 10 90% (0.10) 0% (0.00) 10% (0.10) Close to trunk private/other 5 40% (0.30) 0% (0.00) 60% (0.30) Far from trunk Disp 4 75% (0.25) 25% (0.25) 0% (0.00) Far from trunk church 1 0% () 100% () 0% () Far from trunk government 2 100% (0.00) 0% (0.00) 0% (0.00) Far from trunk private/other 1 100% () 0% () 0% () Very far from trunk Disp 3 33% (0.33) 0% (0.00) 66% (0.33) Very far from trunk church 1 0% () 0% () 100% () Very far from trunk government 2 50% (0.50) 0% (0.00) 50% (0.50) Urban/City 40 80% (0.16) 7% (0.07) 12% (0.11) Major Town 24 66% (0.23) 0% (0.00) 33% (0.23) Close to trunk 18 72% (0.21) 0% (0.00) 27% (0.21) Far from trunk 4 75% (0.25) 25% (0.25) 0% (0.00) Very far from trunk 3 33% (0.33) 0% (0.00) 66% (0.33) Standard errors in parentheses 18  Table 11: Scheduled personnel present and not present by zone level and system: Clinicians and Nurses Zone Level Owner scheduled Present Freq Pres Not Pres Urban/City Disp private/other 4 100% (0.00) 0% (0.00) 0% (0.00) Urban/City HCenter government 41 75% (0.19) 7% (0.07) 17% (0.15) Urban/City Hospital church 21 90% (0.09) 9% (0.09) 0% (0.00) Urban/City Hospital government 6 100% (0.00) 0% (0.00) 0% (0.00) Urban/City Hospital private/other 19 89% (0.10) 0% (0.00) 10% (0.10) Major Town HCenter government 5 60% (0.30) 0% (0.00) 40% (0.30) Major Town Hospital government 57 73% (0.20) 0% (0.00) 26% (0.20) Close to trunk Disp church 7 71% (0.24) 0% (0.00) 28% (0.24) Close to trunk Disp government 17 88% (0.11) 0% (0.00) 11% (0.11) Close to trunk Disp private/other 6 50% (0.30) 0% (0.00) 50% (0.30) Close to trunk HCenter government 6 83% (0.17) 0% (0.00) 16% (0.17) Far from trunk Disp church 2 50% (0.50) 50% (0.50) 0% (0.00) Far from trunk Disp government 3 66% (0.33) 0% (0.00) 33% (0.33) Far from trunk Disp private/other 1 100% () 0% () 0% () Very far from trunk Disp church 1 0% () 0% () 100% () Very far from trunk Disp government 4 50% (0.33) 0% (0.00) 50% (0.33) Urban/City Disp 4 100% (0.00) 0% (0.00) 0% (0.00) Urban/City Hospital 46 91% (0.08) 4% (0.04) 4% (0.04) Urban/City church 21 90% (0.09) 9% (0.09) 0% (0.00) Urban/City government 47 78% (0.17) 6% (0.06) 14% (0.13) Urban/City private/other 23 91% (0.08) 0% (0.00) 8% (0.08) Major Town Hospital 57 73% (0.20) 0% (0.00) 26% (0.20) Major Town government 62 72% (0.20) 0% (0.00) 27% (0.20) Close to trunk Disp 30 76% (0.19) 0% (0.00) 23% (0.19) Close to trunk church 7 71% (0.24) 0% (0.00) 28% (0.24) Close to trunk government 23 86% (0.12) 0% (0.00) 13% (0.12) Close to trunk private/other 6 50% (0.30) 0% (0.00) 50% (0.30) Far from trunk Disp 6 66% (0.27) 16% (0.17) 16% (0.17) Far from trunk church 2 50% (0.50) 50% (0.50) 0% (0.00) Far from trunk government 3 66% (0.33) 0% (0.00) 33% (0.33) Far from trunk private/other 1 100% () 0% () 0% () Very far from trunk Disp 5 40% (0.30) 0% (0.00) 60% (0.30) Very far from trunk church 1 0% () 0% () 100% () Very far from trunk government 4 50% (0.33) 0% (0.00) 50% (0.33) Urban/City 91 84% (0.13) 5% (0.05) 9% (0.09) Major Town 62 72% (0.20) 0% (0.00) 27% (0.20) Close to trunk 36 77% (0.18) 0% (0.00) 22% (0.18) Far from trunk 6 66% (0.27) 16% (0.17) 16% (0.17) Very far from trunk 5 40% (0.30) 0% (0.00) 60% (0.30) Standard errors in parentheses 19  Table 12: Regression analysis of present and not present personnel % Present % Present % Not Present % Not Present Coef. Std. Err Coef. Std. Err Coef. Std. Err Coef. Std. Err major town -0.12 0.08 -0.42 0.16 * 0.17 0.08 * 0.43 0.16 * close to trunk -0.09 0.16 -0.33 0.19 * 0.14 0.15 0.35 0.19 * far from trunk -0.19 0.24 -0.41 0.25 0.09 0.24 0.28 0.25 very far from trunk -0.46 0.26 * -0.75 0.28 * 0.52 0.26 * 0.78 0.28 * dispensary 0.03 0.16 0.07 0.15 -0.02 0.15 -0.06 0.15 government -0.01 0.10 0.25 0.15 -0.09 0.11 -0.26 0.15 * church 0.00 0.12 0.02 0.11 -0.09 0.11 -0.11 0.11 clinician -0.09 0.06 -0.07 0.06 0.06 0.06 0.05 0.06 gov * urban -0.35 0.16 * 0.30 0.16 * constant 0.89 0.09 * 0.93 0.09 * 0.11 0.09 * 0.08 0.09 Adj R-square 2% 9% 7% 12% * significant at the 90% level for a two sided test. Table 12 shows a regression in which we control for the impact of the different variables simultaneously. We show regressions on the percentage of personnel who are present as well as a regression on the percentage of personnel who are not present. The results are very similar, with one important difference. In addition we include the interaction term of government and urban to test whether or not there is a difference in the absenteeism of government personnel in the urban compared to the rural areas. The idea behind this variable is the following: the government is restricted in its ability to hire and fire personnel and therefore can only present low powered incentives to its personnel. On the other hand church services can afford to be much stricter with personnel. However high powered incentives are more useful with frequent monitoring. In the rural areas there is less monitoring but both the government and the church services. Thus, if there is going to be a strong difference between government and church services it should show up in the urban areas. There should be difference in the rural areas, but it would not be as strong. Distance is clearly an important indicator in all four regressions. Being very far from a main road decreases the probability of presence and increases the probability of absence in all four regressions. In addition being in a major town has a smaller but significant impact. This could lend weight to the argument that issues with absenteeism are not simply due to difficulty in reaching the post, but also do to the availability of nearby distractions. The interaction term for the government and the urban area is significant and has sign such that absen- teeism is a greater issue for the government in the urban area. In the last regression the government is seen to be less likely overall to experience absenteeism, but more likely to experience it in the urban areas. 20  3 Physical Infrastructure and Equipment Availability Infrastructure was evaluated using the survey form shown in Figure 8. These scores are shown according to district for government facilities (Table 13), level (Table 14) and owner (Table 15). Table 13: Facility Infrastructure by District Arumeru Arusha Monduli Waiting Rooms room obs 13 10 20 wait room 1 1 1 sit 1 1 .86 good cond 1 1 .93 ventilated .75 .88 1 facility obs 8 4 14 Physical Infrastructure nurse room .13 1 .29 general cond .13 1 .29 vent/lit .13 1 .29 inj room .63 1 .71 rest room .75 .75 .79 latrine .88 1 1 lat cond .5 .75 .86 water .5 .75 .5 Equipment Availability scale .88 .75 .79 height .63 .75 .79 ors mat .5 .75 .46 syr & ndle 1 .75 1 sterilize 1 1 1 antiseptic 1 1 1 bandages 1 1 1 plaster 1 1 1 scissors .75 1 .86 forceps .63 1 .79 sutures .88 1 1 needle hld .5 1 .71 envelopes 1 .75 1 microscope .25 .75 .43 Totals physical inf 3.63 7.25 4.71 equipment 11 12.5 12 total inf 14.63 19.75 16.5 Building paint: poor .38 0 .14 paint: accpt .63 .5 .79 paint: excel 0 .5 .07 roof: poor .25 .25 .14 roof: accpt .75 0 .79 roof: excel 0 .75 .07 grds: poor .25 0 .14 grds: accpt .75 1 .86 grds: excel 0 0 0 Reported is number of facilities observed and fraction of all facilities that scored 1 for present, or acceptable. 21  Table 14: Facility Infrastructure by Level Dispensary Health Center Hospital Waiting Rooms room obs 43 9 11 wait room .97 1 1 sit .84 1 1 good cond .97 1 1 ventilated .94 .9 1 Physical Infrastructure facility obs 31 5 8 nurse room .1 .8 1 general cond .1 .8 1 vent/lit .1 .8 1 inj room .65 1 1 rest room .71 1 1 latrine .94 1 1 lat cond .71 1 1 water .53 .6 .88 Equipment Availability scale .83 .8 1 height .7 .8 1 ors mat .4 1 .63 syr & ndle 1 .8 1 sterilize .97 1 1 antiseptic .97 1 1 bandages 1 1 1 plaster .97 1 1 scissors .9 .6 1 forceps .74 1 1 sutures .94 1 1 needle hld .71 .8 1 envelopes .9 1 .88 microscope .35 1 1 Totals physical inf 3.8 7 7.88 equipment 11.41 12.75 13.5 total inf 15.14 19.5 21.38 Building paint: poor .13 0 .13 paint: accpt .84 .8 .63 paint: excel .03 .2 .25 roof: poor .1 .2 .25 roof: accpt .84 .6 .5 roof: excel .06 .2 .25 grds: poor .06 0 .25 grds: accpt .94 1 .63 grds: excel 0 0 .13 Reported is number of facilities observed and fraction of all facilities that scored 1 for present, or acceptable. 22  Table 15: Facility Infrastructure by Owner COGI Govt Islamic Luth RC SDA Waiting Rooms room obs 3 43 0 9 3 4 wait room 1 1 1 1 1 sit 1 .92 .83 .67 1 good cond 1 .96 1 1 1 ventilated 1 .9 1 1 1 Physical Infrastructure facility obs 3 26 1 6 3 3 nurse room 0 .35 1 .17 .33 .67 general cond 0 .35 1 .17 .33 .67 vent/lit 0 .35 1 .17 .33 .67 inj room .67 .73 1 .67 1 1 rest room .67 .77 1 1 1 .67 latrine 1 .96 1 .83 1 1 lat cond 1 .73 1 .83 .67 1 water 0 .54 1 1 1 .67 Equipment Availability scale 1 .81 1 1 1 1 height 0 .73 1 1 1 1 ors mat .5 .52 1 .5 .67 .33 syr & ndle 1 .96 1 1 1 1 sterilize 1 1 1 1 1 1 antiseptic 1 1 1 1 1 1 bandages 1 1 1 1 1 1 plaster 1 1 1 1 1 1 scissors 1 .85 1 1 1 1 forceps .67 .77 1 1 1 1 sutures 1 .96 1 1 1 1 needle hld .67 .69 1 1 1 1 envelopes 1 .96 1 .83 1 .67 microscope .67 .42 1 .67 .67 1 Totals physical inf 3.33 4.77 8 4.8 5.67 6.33 equipment 11 11.75 14 13 13.33 13 total inf 14.5 16.42 22 17.6 19 19.33 Building paint: poor 0 .19 0 0 0 0 paint: accpt 1 .69 1 1 1 1 paint: excel 0 .12 0 0 0 0 roof: poor .33 .19 0 0 0 0 roof: accpt .67 .65 1 1 1 1 roof: excel 0 .15 0 0 0 0 grds: poor 0 .15 0 0 0 0 grds: accpt 1 .85 1 1 1 1 grds: excel 0 0 0 0 0 0 Reported is number of facilities observed and fraction of all facilities that scored 1 for present, or acceptable. 23  4 Pharmaceutical Availability Each facility was evaluated for the presence of pharmaceutical supplies that would be expected at a dispensary (shown in Figure 9). Shown here is the breakdown according to district for government facilities (Table 16), owner (Table 18) and level (Table 17). The number shown is the fraction of facilities for which that drug is present as well as a total score showing the percentage of all drugs that is present for each type of facility. deliv days is the number of days since the last delivery of drugs. Table 19 shows the presence of drugs for the 4 quartiles of facilities by days since the last delivery, where the first quartile represents less time passed between our evaluation and the last delivery and the fourth quartile represents facilities that have seen the most time pass since the last delivery. Table 16: Pharmaceutical Availability Evaluation by District Arumeru Arusha Monduli med obs 8 4 14 litpres 1 1 .86 sp 1 1 .93 amodiaquin .88 .75 .93 quin inj .25 .75 .29 quin tab .25 1 .5 asa tab 1 1 .93 pcm 1 1 .93 ors 1 1 1 cotri tab 1 1 .93 cotri syr 1 1 1 pen g 1 1 .93 pen v 1 1 .93 amp tab 0 0 .21 amp syr 0 0 .21 tetra .38 .25 .5 metronidaz 1 1 1 mebendazol 1 1 1 tetra eye 1 1 .86 bbe 1 1 .93 multi vit .13 .25 .07 drug supply .73 .79 .74 deliv days 21.5 16.3 16.3 Reported is number of facilities observed and fraction of all facilities that scored 1 for having the drug present. 24  Table 17: Pharmaceutical Availability Evaluation by Level Dispensary Health Center Hospital med obs 31 5 8 litpres .84 1 1 sp 1 1 .88 amodiaquin .84 1 .5 quin inj .42 .4 1 quin tab .55 .8 1 asa tab .97 1 1 pcm .97 1 1 ors .94 1 1 cotri tab .97 1 1 cotri syr 1 1 1 pen g .97 1 1 pen v .84 1 1 amp tab .32 .2 .5 amp syr .39 .2 .5 tetra .45 .4 1 metronidaz .97 1 1 mebendazol 1 1 1 tetra eye .9 1 .88 bbe .97 1 1 multi vit .42 0 .5 drug supply .78 .79 .88 deliv days 18.5 18.6 16.5 Reported is number of facilities observed and fraction of all facilities that scored 1 for having the drug present. Table 18: Pharmaceutical Availability Evaluation by Owner COGI Govt Islamic Luth RC SDA med obs 3 26 1 6 3 3 litpres .67 .92 1 1 .67 1 sp 1 .96 1 1 1 1 amodiaquin .67 .88 1 .83 0 1 quin inj 1 .35 1 1 .67 .33 quin tab 1 .5 1 1 1 .67 asa tab 1 .96 1 1 1 1 pcm 1 .96 1 1 1 1 ors 1 1 1 .83 1 1 cotri tab 1 .96 1 1 1 1 cotri syr 1 1 1 1 1 1 pen g 1 .96 1 1 1 1 pen v 1 .96 1 .67 .67 1 amp tab 1 .12 0 .67 .33 1 amp syr 1 .12 0 .83 .67 1 tetra 1 .42 1 .67 .33 .67 metronidaz .67 1 1 1 1 1 mebendazol 1 1 1 1 1 1 tetra eye 1 .92 1 .83 1 1 bbe 1 .96 1 1 1 1 multi vit 1 .12 1 .67 .67 1 drug supply .96 .74 .89 .89 .81 .93 deliv days 3 17.9 8 24 30.7 22 Reported is number of facilities observed and fraction of all facilities that scored 1 for having the drug present. 25  Table 19: Pharmaceutical Availability by Quartile of Last Delivery of Drugs Days since Last Delivery Quartile first second third fourth med obs 11 12 10 11 litpres .82 .83 1 .91 sp .91 1 1 1 amodiaquin .64 .92 1 .64 quin inj .73 .42 .3 .64 quin tab .91 .5 .4 .82 asa tab 1 1 .9 1 pcm 1 1 .9 1 ors 1 .92 .9 1 cotri tab 1 1 .9 1 cotri syr 1 1 1 1 pen g .91 1 1 1 pen v 1 .92 .8 .82 amp tab .55 .25 .2 .36 amp syr .55 .25 .3 .45 tetra .82 .5 .4 .45 metronidaz .91 1 1 1 mebendazol 1 1 1 1 tetra eye .91 .92 1 .82 bbe 1 1 1 .91 multi vit .64 .25 .2 .45 drug supply .87 .78 .75 .81 Reported is number of facilities observed and fraction of all facilities that scored 1 for having the drug present. 26  5 Examination Room Evaluation and Clinician Qualifications As well as the general equipment available in each facility we evaluated the presence of necessary equipment in the consultation room. As shown in Figure 2, this is a list of equipment that would be necessary to evaluate or diagnose the range of illnesses that should be expected at a dispensary. In addition, we show here some basic statistics about the qualifications of the clinician evaluated. The qualifications and consultation room equipment are shown broken down by district (Table 20), owner (Table 22) and level (Table 21). Table 20: Clinician Qualification and Examining Room Characteristics by District Arumeru Arusha Monduli Clinician characterisitcs clinicn obs 9 11 18 doctor 0 .27 .06 officer .33 .45 .33 assistant .44 .18 .44 nurse .22 0 .17 experience 13.56 17.8 15.33 tenure 4.02 10.83 9.69 Examining room table 1 1 1 bed .67 .8 .88 wash basin .22 .8 .44 stetho 1 .9 1 thermo .89 1 1 bp machine .33 .8 .94 otoscope .11 .2 0 spatula .33 .5 .25 torch 0 0 .06 gloves .78 .6 .88 cards .89 .8 .88 well lit .67 .9 1 Reported is number of clinicians or consulting rooms seen and the fraction of the clinicians with a given characteristic or of consulting rooms with a given feature present. 27  Table 21: Clinician Qualification and Examining Room Characteristics by Level Dispensary Health Center Hospital Clinician characterisitcs clinicn obs 38 8 23 doctor .08 .25 .26 officer .34 .5 .61 assistant .34 .25 .04 nurse .21 0 0 experience 14.43 15.5 15.43 tenure 6.59 2.67 9.81 Examining room table .97 1 1 bed .81 .86 1 wash basin .49 1 .84 stetho 1 .86 .95 thermo 1 1 .89 bp machine .84 .86 .84 otoscope .16 .14 .53 spatula .38 .29 .63 torch .32 0 .47 gloves .95 .71 .68 cards .92 .71 .79 well lit .89 .86 1 Reported is number of clinicians or consulting rooms seen and the fraction of the clinicians with a given characteristic or of consulting rooms with a given feature present. Table 22: Clinician Qualification and Examining Room Characteristics by Owner COGI Govt Islamic Luth RC SDA Clinician characterisitcs clinicn obs 3 38 3 12 5 4 doctor .67 .11 .67 0 0 0 officer .33 .37 .33 .75 .4 .75 assistant 0 .37 0 .08 .2 0 nurse 0 .13 0 .17 .2 0 experience 21 15.57 19.33 9.75 14 10.67 tenure 5.21 8.62 12.44 2.74 3.46 2.31 Examining room table 1 1 1 1 1 1 bed 1 .8 1 1 1 1 wash basin 1 .49 1 .73 1 1 stetho 1 .97 1 1 1 1 thermo 1 .97 .67 1 1 1 bp machine 1 .74 1 1 1 1 otoscope 0 .09 1 .55 .5 .33 spatula .33 .34 1 .55 .25 1 torch 1 .03 1 .64 .5 .67 gloves 1 .77 1 1 .75 1 cards 1 .86 1 .91 .5 1 well lit 1 .89 1 1 1 1 Reported is number of clinicians or consulting rooms seen and the fraction of the clinicians with a given characteristic or of consulting rooms with a given feature present. 28  6 Nursing Quality (Dispensing, Injections and Dressing) Nursing quality was evaluated for quality in the dispensing of drugs (see Figure 6), giving injections and dressing wounds (see Figure 7). There were different numbers of observations for these three procedures. We saw many drug dispensing procedures and considerably fewer wound dressings. The results are shown by district (Table 23), owner (Table 25) and level (Table 24). Table 23: Nursing Quality (Dispensing, Injections and Dressing) by District Arumeru Arusha Monduli drug obs 139 109 182 dispense 1 .99 .98 label .68 .95 .84 explain .93 .88 .96 side eff .02 .35 .05 complet .02 .28 .1 understood .78 .73 .79 polite .95 .85 .98 wash hand .36 .92 .38 injctn obs 64 25 55 load pres 1 1 .96 load corr .94 1 .89 give inj .88 1 .85 sterile .97 1 1 disassem .55 .12 .16 wound obs 9 6 10 technique 0 .5 .5 clean .22 .83 .5 dress .44 1 .9 Reported is number of observations of drug dispensing, injections and wound dressing as well as the fraction of each observation that correctly followed the given procedure. 29  Table 24: Nursing Quality (Dispensing, Injections and Dressing) by Facility Level Dispensary Health Center Hospital drug obs 259 144 125 dispense .98 1 .98 label .67 .94 .95 explain .96 .9 .92 side eff .05 .22 .09 complet .09 .2 .09 understood .76 .83 .79 polite .98 .9 .95 wash hand .4 .56 .64 injctn obs 127 35 20 load pres .98 1 1 load corr .95 .89 .9 give inj .87 .97 1 sterile .98 1 .95 disassem .39 .09 .3 wound obs 21 7 17 technique .24 .43 .59 clean .33 .71 .82 dress .57 1 .82 Reported is number of observations of drug dispensing, injections and wound dressing as well as the fraction of each observation that correctly followed the given procedure. Table 25: Nursing Quality (Dispensing, Injections and Dressing) by Facility Owner COGI Govt Islamic Luth RC SDA drug obs 24 430 11 34 15 5 dispense 1 .99 .82 1 1 1 label .33 .82 1 .91 .87 1 explain 1 .93 .82 .97 .8 1 side eff .05 .11 .2 .09 .15 0 complet .08 .12 0 .03 .36 .5 understood .88 .77 .64 1 .8 1 polite 1 .94 1 1 1 1 wash hand 1 .48 0 .61 .53 .4 injctn obs 9 144 4 12 4 2 load pres 1 .99 1 1 1 .5 load corr 1 .93 .75 1 .75 1 give inj 1 .89 1 1 1 1 sterile 1 .99 1 1 1 1 disassem .75 .33 0 0 .25 0 wound obs 2 25 4 8 3 1 technique 1 .32 .25 .63 0 0 clean 1 .48 .25 .75 1 0 dress 1 .76 .25 .75 1 0 Reported is number of observations of drug dispensing, injections and wound dressing as well as the fraction of each observation that correctly followed the given procedure. 30  7 Consultation Quality Clinicians from our team sat in on the consultations that were done by the clinicians being observed. In each consultation they observed what was done by the clinician and compared that to a checklist (see Figure 3, 4, and 5). For many of the items it is straightforward to judge whether or not the clinician performed the activity and therefore to evaluate the quality with which he performs services. However, each patient is different and there are times when a clinician should examine the patient for certain things and times when he does not need to. To get around this problem we developed 4 conditions that we expected to observe very closely (fever, cough, diarrhea, and symptoms indicative of STDs). For these conditions is was more clear what the clinician should do, and what he was not expected to do. For any of these 4 conditions, then, we filled a more detailed check list. The results are divided into three sets of tables. In the first set of tables (examined by district for government facilities (Table 26), owner (Table 26), level (Table 29), cadre (Table 35), quartiles of experience (Table 38)) and both cadre and quartile of experience (Table 41)) we show the score (where 1 indicates they did the activity listed, 0 that they did not, and any score between 0 and 1 is the fraction of observations observed for which we observed the clinician as performing that activity), on welcoming, general history taking and closing as well as the scores for illnesses with fever as a major symptom. In the second set of tables (examined by district for government facilities (Table 27), owner (Table 27), level (Table 30), cadre (Table 36), quartiles of experience (Table 39) and both cadre and quartile of experience (Table 42)) we show the results for cough, diarrhea, symptoms indicative of STDs and general conditions. In the third set of tables (examined by district for government facilities (Table 28), owner (Table 28), level (Table 31), cadre (Table 37), quartiles of experience (Table 40) and both cadre and quartile of experience (Table 43)) we show overall scores in the categories of opening the consultation, explaining the diagnosis and closing the consultation, history taking, physical examination and a combination of history taking and physical examination. The quartiles of experience go from the least experience to the most. To achieve each overall score we calculated the score as a percent of the total possible score1. To compare fever, cough and diarrhea (we drop STD and general in the total score) we normalize each final score so that the scores across each illness type have a mean of 0 and a standard deviation of 1. The we take the total across all three presenting symptoms. This method allows us to compare the quality of a fever consultation with that of a diarrhea consultation. 1Taking into account conditions for which a particular question or procedure is not applicable (asking about vaccination history for a 25 year old with a cough, for example). 31  Table 26: Consultation Quality by District: Details (part I) Arumeru Arusha Monduli Receiving Patients begin cons 176 180 220 welcome .33 .83 .56 greet .28 .68 .51 look at .86 .88 .83 chair .97 .88 .95 Basic History Taking cons 136 170 194 symp duration .79 .94 .82 probe deeper .65 .54 .76 other symptoms .54 .53 .61 o sympt duration .09 .53 .25 othr treatment .24 .47 .39 Ending the Consultation close cons 131 154 181 tell diagnosis .05 .55 .16 explain diagnosis .16 .77 .25 explain treatment .5 .81 .59 health education .21 .54 .27 discuss return .14 .51 .2 listen .93 .98 .92 let talk .9 .97 .93 ensure undersd .64 .48 .34 Fever fever obs 27 72 59 fever pattern .82 .89 .92 chills sweats .14 .14 .15 cough in fever .43 .51 .41 diarrhea/vomit .14 .59 .29 convulsions .04 0 .04 fever: gen exam .59 .69 .58 temp for fever .36 .71 .59 anemia .25 .44 .25 ear/throat .25 .44 .24 palpate spleen .18 .11 .03 blood slide .14 .19 .22 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 32  Table 27: Consultation Quality by District: Details (part II) Arumeru Arusha Monduli Cough cough obs 59 54 68 duration of cough .81 .94 .79 sputum prod .29 .45 .39 blood sput .03 .08 .09 chest pain .12 .31 .38 diff breathing .09 .42 .13 vaccinations .04 .09 .11 fever w/ cough .58 0 .31 cough: gen exam .25 .69 .53 resp rate .05 .43 .04 indrawing .05 .5 .1 throat .14 .37 .19 auscultate .81 .37 .68 temp for cough .16 0 .19 Diarr diarrhea obs 8 14 16 frequency .75 .86 .56 consistency .5 .64 .63 blood/mucus .13 .54 .44 vomiting .38 .64 .31 fever w/ diarr .63 .64 .5 diarr: gen exam .5 .5 .5 lethargy .5 .71 .31 fontanella .14 .14 .25 pinch skin .14 .54 0 weigh .14 .79 .2 temp for diarr .13 .14 .13 STD std obs 1 16 5 discharge/ulcer 1 .75 .6 pain or itching 1 .4 .4 fever w/ std 0 .19 0 pain on urination 0 .5 .4 sexual history 0 .13 0 prev exposure 0 0 0 partners 0 .18 0 std: gen exam 0 .5 .4 skin rash 0 .13 .2 lymph nodes 0 .25 0 tenderness 1 .6 .8 genitalia 1 .5 .6 precautions 0 .2 0 Other general obs 41 45 57 history to sympt .96 .88 .94 other: gen exam .46 .49 .26 exam to sympts .59 .75 .57 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 33  Table 28: Consultation Quality by District: Total Scores Arumeru Arusha Monduli summary obs 89 128 139 o: open (avg) -.48 .21 -.14 o: expl (avg) -.61 .63 -.38 o: close (avg) .19 .07 -.07 o: score (avg) -.39 .25 -.24 o: hist (avg) -.13 .2 0 o: exam (avg) -.46 .16 -.33 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). Table 29: Consultation Quality by Facility Level: Details (part I) Dispensary Health Center Hospital Receiving Patients begin cons 361 172 207 welcome .57 .54 .84 greet .51 .47 .75 look at .86 .78 .97 chair .92 .92 .99 Basic History Taking cons 289 155 191 symp duration .82 .9 .83 probe deeper .58 .73 .69 other symptoms .61 .48 .49 o sympt duration .25 .23 .33 othr treatment .37 .26 .47 Ending the Consultation close cons 279 142 155 tell diagnosis .24 .25 .48 explain diagnosis .34 .49 .57 explain treatment .63 .77 .62 health education .35 .36 .44 discuss return .3 .37 .32 listen .92 .98 .92 let talk .93 .96 .9 ensure undersd .5 .45 .41 Fever fever obs 87 48 72 fever pattern .85 .89 .79 chills sweats .13 .11 .17 cough in fever .45 .49 .4 diarrhea/vomit .38 .47 .24 convulsions .01 .02 .07 fever: gen exam .62 .69 .78 temp for fever .8 .54 .42 anemia .25 .56 .44 ear/throat .28 .52 .47 palpate spleen .05 .15 .33 blood slide .15 .19 .43 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 34  Table 30: Consultation Quality by Facility Level: Details (part II) Dispensary Health Center Hospital Cough cough obs 110 55 56 duration of cough .83 .95 .8 sputum prod .32 .42 .33 blood sput .04 .07 .13 chest pain .31 .3 .33 diff breathing .21 .24 .16 vaccinations .14 0 .07 fever w/ cough .39 .23 .15 cough: gen exam .44 .53 .75 resp rate .17 .16 .11 indrawing .21 .2 .25 throat .2 .31 .48 auscultate .59 .64 .86 temp for cough .28 .02 .04 Diarr frequency .75 .71 .75 consistency .68 .57 .67 blood/mucus .52 .14 .25 vomiting .5 0 .67 fever w/ diarr .61 .29 .67 diarr: gen exam .54 .29 .75 lethargy .57 .29 .83 fontanella .19 0 .5 pinch skin .3 0 .1 weigh .54 .14 .42 temp for diarr .19 0 .17 STD diarrhea obs 28 7 12 std obs 8 13 6 discharge/ulcer .63 .77 .5 pain or itching .63 .33 .33 fever w/ std 0 .23 0 pain on urination .63 .31 .67 sexual history 0 .08 .33 prev exposure 0 0 0 partners .13 .11 0 std: gen exam .13 .62 .67 skin rash 0 .23 0 lymph nodes .13 .15 .33 tenderness .5 .75 .67 genitalia .38 .62 .17 precautions .25 .08 .17 Other general obs 80 40 60 history to sympt .96 .84 .91 other: gen exam .39 .38 .47 exam to sympts .66 .71 .64 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 35  Table 31: Consultation Quality by Facility Level: Total Scores Dispensary Health Center Hospital summary obs 210 106 138 o: open (avg) -.13 -.25 .45 o: expl (avg) -.14 .08 .18 o: close (avg) .09 .02 -.17 o: score (avg) -.03 -.04 .08 o: hist (avg) .09 .01 -.15 o: exam (avg) -.15 -.09 .3 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). Table 32: Consultation Quality by Facility Owner: Details (part I) COGI Govt Islamic Luth RC SDA Receiving Patients begin cons 32 576 25 49 26 6 welcome .88 .57 .92 .78 .81 1 greet .88 .49 .92 .9 .54 1 look at .88 .86 .96 .98 .96 1 chair .88 .93 .96 1 .96 1 Basic History Taking cons 25 500 25 33 25 5 symp duration .88 .85 .88 .88 .88 1 probe deeper .48 .66 .4 .95 .5 1 other symptoms .84 .56 .64 .4 .13 .4 o sympt duration .12 .29 .4 .13 .29 0 othr treatment .36 .38 .16 .54 .36 .6 Ending the Consultation close cons 23 466 17 28 16 4 tell diagnosis .65 .26 .76 .29 .25 .25 explain diagnosis .57 .4 .59 .64 .44 .75 explain treatment .87 .64 .76 .93 .69 1 health education .65 .34 .71 .61 .13 .5 discuss return .57 .29 .71 .36 .13 1 listen 1 .94 .83 1 .94 1 let talk 1 .93 .83 1 .94 1 ensure undersd .43 .48 .39 .47 .18 .5 Fever fever obs 10 158 18 4 10 1 fever pattern .8 .89 .33 1 1 1 chills sweats 0 .15 .11 .25 0 0 cough in fever .3 .46 .33 0 .5 0 diarrhea/vomit .1 .4 .17 0 .3 0 convulsions 0 .02 0 0 .44 0 fever: gen exam .9 .63 1 .75 .9 0 temp for fever .8 .6 .28 .75 .9 0 anemia .4 .34 .89 .5 .3 0 ear/throat .5 .33 .89 .25 .4 1 palpate spleen .2 .09 .83 0 .2 0 blood slide .6 .19 .17 .75 .5 0 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 36  Table 33: Consultation Quality by Facility Owner: Details (part II) COGI Govt Islamic Luth RC SDA Cough cough obs 7 181 12 8 7 0 duration of cough 1 .84 .83 .88 1 sputum prod .2 .38 .17 .5 .14 blood sput .2 .07 0 .17 .2 chest pain .4 .28 .42 .5 .14 diff breathing .43 .2 .33 .13 .14 vaccinations .5 .08 .17 0 0 fever w/ cough .71 .3 0 .25 .14 cough: gen exam .86 .49 1 .38 1 resp rate 0 .16 .33 0 0 indrawing .29 .21 .58 .25 0 throat .86 .23 .83 .13 .71 auscultate .86 .63 .83 .88 .86 temp for cough .86 .12 0 .38 .14 Diarr diarrhea obs 0 38 2 2 2 1 frequency .71 .5 1 1 1 consistency .61 .5 1 1 1 blood/mucus .41 .5 0 .5 0 vomiting .45 .5 .5 1 0 fever w/ diarr .58 .5 .5 1 1 diarr: gen exam .5 1 .5 1 1 lethargy .5 1 1 1 1 fontanella .19 .5 .5 1 0 pinch skin .22 0 1 0 0 weigh .42 0 0 1 1 temp for diarr .14 0 .5 .5 0 STD std obs 0 22 1 3 1 0 discharge/ulcer .73 0 .33 1 pain or itching .43 0 .33 1 fever w/ std .14 0 0 0 pain on urination .45 0 1 0 sexual history .1 0 .33 0 prev exposure 0 0 0 0 partners .12 0 0 0 std: gen exam .45 1 .33 1 skin rash .14 0 0 0 lymph nodes .18 1 0 0 tenderness .67 1 .33 1 genitalia .55 0 0 0 precautions .14 0 .33 0 Other general obs 9 143 5 14 1 3 history to sympt .96 .92 .83 1 1 1 other: gen exam .56 .39 .8 .43 1 .33 exam to sympts .79 .65 .87 .35 .79 .4 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 37  Table 34: Consultation Quality by Facility Owner: Total Scores COGI Govt Islamic Luth RC SDA summary obs 17 356 32 14 19 2 o: open (avg) .41 -.14 .63 .54 .21 .83 o: expl (avg) .7 -.11 .84 .42 -.28 .82 o: close (avg) .08 .06 -.42 .1 -.35 .17 o: score (avg) .65 -.1 .31 .07 .54 -.49 o: hist (avg) .06 .04 -.66 -.04 .25 -.37 o: exam (avg) .9 -.18 1.06 .16 .65 -.26 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). Table 35: Consultation Quality by Cadre: Details (part I) doctor officer assist nurse Receiving Patients begin cons 76 203 329 132 welcome .33 .58 .62 .94 greet .3 .48 .6 .8 look at .89 .84 .86 .95 chair .99 .9 .94 .97 Basic History Taking cons 57 176 281 121 symp duration .79 .82 .88 .83 probe deeper .63 .55 .72 .56 other symptoms .38 .5 .6 .53 o sympt duration .09 .17 .3 .4 othr treatment .23 .31 .45 .36 Ending the Consultation close cons 57 164 247 108 tell diagnosis .05 .19 .37 .49 explain diagnosis .05 .37 .48 .65 explain treatment .54 .55 .75 .68 health education .16 .26 .46 .5 discuss return .14 .23 .36 .46 listen .92 .95 .95 .91 let talk .93 .94 .94 .89 ensure undersd .49 .37 .52 .43 Fever fever obs 12 47 95 53 fever pattern .67 .96 .92 .63 chills sweats .17 .11 .18 .08 cough in fever .33 .38 .48 .45 diarrhea/vomit .08 .45 .31 .38 convulsions 0 .02 .07 0 fever: gen exam .33 .79 .64 .77 temp for fever .5 .81 .53 .58 anemia .25 .23 .39 .57 ear/throat .25 .28 .4 .57 palpate spleen 0 .19 .1 .3 blood slide 0 .17 .3 .3 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 38  Table 36: Consultation Quality by Cadre: Details (part II) doctor officer assist nurse Cough cough obs 27 58 98 38 duration of cough .77 .84 .89 .84 sputum prod .2 .35 .38 .35 blood sput 0 .1 .08 .07 chest pain .4 .25 .28 .43 diff breathing .07 .1 .33 .14 vaccinations .19 .03 .07 .12 fever w/ cough .59 .16 .35 .11 cough: gen exam .22 .53 .54 .76 resp rate 0 .12 .21 .16 indrawing .04 .14 .28 .32 throat .11 .21 .31 .55 auscultate .59 .71 .65 .71 temp for cough .33 .21 .07 .13 Diarr diarrhea obs 4 15 22 6 frequency .5 .87 .73 .67 consistency .75 .67 .64 .67 blood/mucus .5 .33 .43 .33 vomiting 0 .53 .5 .5 fever w/ diarr 0 .6 .68 .5 diarr: gen exam .25 .67 .59 .33 lethargy .25 .53 .68 .67 fontanella .25 .14 .32 .17 pinch skin 0 .23 .25 .17 weigh 0 .5 .45 .5 temp for diarr .33 .13 .18 0 STD std obs 2 6 12 7 discharge/ulcer 0 .83 .75 .57 pain or itching .5 .2 .5 .43 fever w/ std 0 .33 0 .14 pain on urination .5 .5 .58 .29 sexual history 0 .2 .17 0 prev exposure 0 0 0 0 partners 0 0 .18 0 std: gen exam 0 .5 .5 .57 skin rash 0 .33 .08 0 lymph nodes 0 0 .17 .43 tenderness .5 .6 .75 .57 genitalia .5 .67 .33 .43 precautions 0 .2 .17 .14 Other general obs 19 51 76 34 history to sympt 1 .9 .92 .9 other: gen exam .05 .57 .43 .32 exam to sympts .5 .64 .66 .77 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 39  Table 37: Consultation Quality by Cadre: Total Scores nurse assistant officer doctor summary obs 42 117 202 93 o: open (avg) -.43 -.18 0 .53 o: expl (avg) -.68 -.3 .19 .39 o: close (avg) .09 -.04 .08 -.17 o: score (avg) -.53 -.07 .08 .15 o: hist (avg) -.17 -.04 .16 -.22 o: exam (avg) -.63 -.06 -.01 .39 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). Table 38: Consultation Quality by Quartile of Experience: Details (part I) Quartile first second third fourth Receiving Patients begin cons 139 176 181 244 welcome .47 .62 .76 .66 greet .51 .46 .72 .57 look at .81 .96 .87 .84 chair .96 .95 .92 .94 Basic History Taking cons 117 143 161 214 symp duration .75 .9 .93 .79 probe deeper .54 .93 .65 .55 other symptoms .58 .55 .71 .38 o sympt duration .12 .34 .41 .2 othr treatment .28 .51 .42 .3 Ending the Consultation close cons 113 133 140 190 tell diagnosis .13 .28 .46 .33 explain diagnosis .22 .42 .57 .48 explain treatment .65 .65 .74 .62 health education .22 .45 .46 .36 discuss return .19 .3 .48 .29 listen .93 .95 .95 .92 let talk .9 .94 .94 .93 ensure undersd .63 .45 .45 .36 Fever fever obs 19 39 74 75 fever pattern .84 .82 .78 .91 chills sweats .11 .21 .12 .12 cough in fever .37 .28 .52 .47 diarrhea/vomit .21 .17 .41 .41 convulsions .08 0 .01 .07 fever: gen exam .37 .67 .74 .73 temp for fever .47 .35 .74 .64 anemia .21 .2 .42 .51 ear/throat .37 .2 .46 .47 palpate spleen .11 .1 .2 .19 blood slide .3 .28 .18 .31 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 40  Table 39: Consultation Quality by Quartile of Experience: Details (part II) first second third fourth Cough cough obs 44 49 57 71 duration of cough .84 .94 .88 .79 sputum prod .38 .51 .31 .25 blood sput .06 0 .13 .07 chest pain .29 .27 .34 .32 diff breathing .07 .13 .46 .14 vaccinations 0 .03 .19 .08 fever w/ cough .6 .28 .16 .22 cough: gen exam .3 .53 .74 .54 resp rate .02 .15 .32 .1 indrawing .02 .21 .4 .2 throat .14 .22 .42 .35 auscultate .77 .61 .54 .75 temp for cough .25 .06 .21 .1 Diarr diarrhea obs 9 6 16 16 frequency .78 1 .69 .69 consistency .78 .67 .63 .63 blood/mucus .56 .33 .53 .19 vomiting .11 .5 .75 .38 fever w/ diarr .33 .83 .75 .44 diarr: gen exam .33 .83 .69 .44 lethargy .22 .83 .69 .63 fontanella .13 .5 .25 .19 pinch skin 0 .4 .31 .15 weigh .14 .33 .63 .44 temp for diarr .13 .17 .25 .06 STD std obs 2 4 8 13 discharge/ulcer 1 .75 .63 .62 pain or itching 0 .5 .63 .33 fever w/ std 0 0 0 .23 pain on urination 0 .5 .38 .62 sexual history 0 .25 0 .17 prev exposure 0 0 0 0 partners 0 0 .29 0 std: gen exam .5 .25 .38 .62 skin rash .5 0 0 .15 lymph nodes 0 .5 .38 0 tenderness 1 1 .75 .42 genitalia 1 .75 .25 .38 precautions 0 .5 .13 .08 Other general obs 40 53 42 45 history to sympt .96 .96 .89 .89 other: gen exam .5 .49 .24 .4 exam to sympts .65 .61 .84 .57 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 41  Table 40: Consultation Quality by Quartile of Experience: Total Scores first second third fourth summary obs 70 86 139 159 o: open (avg) -.23 -.02 .21 -.01 o: expl (avg) -.4 0 .35 -.02 o: close (avg) .11 .1 .04 -.17 o: score (avg) -.31 -.29 .31 .02 o: hist (avg) 0 -.14 .15 -.05 o: exam (avg) -.47 -.27 .3 .09 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). Table 41: Consultation Quality by Quartile of Experience and Cadre: Details (part I) Cadre Doctor C. Officer C. Assist Nurse Exp. Quartile 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Receiving Patients begin cons 12 19 5 40 38 51 34 80 80 91 81 77 9 15 61 47 welcome .83 .05 .8 .25 .63 .57 .41 .64 .29 .7 .79 .69 .89 1 .9 .98 greet .67 .05 1 .22 .58 .24 .56 .56 .41 .6 .73 .64 .89 .87 .79 .77 look at 1 1 1 .8 .87 .88 .85 .8 .76 .99 .81 .84 .78 1 .95 .96 chair 1 1 1 .98 .89 .86 1 .89 .98 .99 .83 .97 1 1 .98 .94 Basic History Taking cons 12 8 5 32 37 43 27 69 60 78 73 70 8 14 56 43 symp duration .75 1 .8 .75 .54 .88 .93 .9 .9 .92 .96 .71 .63 .79 .91 .77 probe deeper .5 1 .8 .5 .07 .93 .93 .6 .75 .98 .56 .69 .63 .67 .7 .28 other symptoms .09 .7 .2 .41 .59 .26 .74 .51 .67 .61 .77 .36 .5 1 .67 .21 o sympt duration .33 .11 0 .07 .03 .36 .09 .18 .12 .27 .45 .31 .38 .71 .46 .18 othr treatment .27 .13 .2 .25 .35 .3 .48 .23 .25 .67 .41 .41 .25 .5 .41 .28 Ending the Consultation close cons 12 10 5 30 34 39 25 66 60 69 63 55 7 15 47 39 tell diagnosis .08 0 .2 .03 .06 .21 .04 .3 .08 .29 .68 .42 1 .6 .4 .46 explain diagnosis .08 0 .2 .03 .12 .44 .2 .52 .22 .42 .76 .53 1 .67 .55 .69 explain treatment .92 .4 1 .37 .32 .62 .44 .68 .78 .68 .9 .64 .71 .73 .66 .67 health education .25 .1 0 .17 .09 .33 0 .39 .27 .49 .71 .33 .43 .8 .4 .51 discuss return .17 .1 .2 .13 .06 .21 .08 .38 .2 .35 .73 .13 .86 .47 .38 .49 listen 1 .9 1 .88 .85 .93 .96 1 .95 .97 .99 .86 1 .94 .9 .9 let talk 1 1 1 .88 .85 .91 .92 1 .9 .97 .99 .9 1 .88 .88 .9 ensure undersd .25 0 .8 .56 .64 .31 .38 .26 .71 .5 .46 .4 .5 .63 .41 .36 Fever fever obs 4 3 0 5 4 11 11 21 10 21 39 25 1 4 24 24 fever pattern .75 0 1 .75 .91 1 1 1 .95 .9 .88 0 .5 .48 .83 chills sweats .25 0 .2 0 .27 0 .1 .1 .18 .24 .12 0 .5 0 .13 cough in fever .25 .67 .2 .5 .36 .64 .24 .4 .23 .59 .56 0 0 .35 .63 diarrhea/vomit 0 0 .2 .5 .18 .45 .57 .1 .18 .51 .2 1 .25 .22 .54 convulsions 0 0 0 0 0 .09 0 .13 0 0 .2 0 0 0 0 fever: gen exam .5 .67 0 .25 .82 .91 .81 .4 .67 .69 .64 0 .25 .75 .92 temp for fever .5 .67 .4 .75 .55 1 .86 .3 .27 .69 .6 1 0 .71 .54 anemia .25 .67 0 0 .09 0 .48 .3 .18 .44 .52 0 .25 .58 .63 ear/throat .25 .67 0 .25 0 .09 .52 .5 .18 .44 .48 0 .5 .67 .5 palpate spleen 0 0 0 .25 .27 0 .24 .1 .05 .1 .16 0 0 .46 .21 blood slide 0 0 0 .4 .18 0 .19 .3 .41 .18 .4 1 0 .25 .38 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 42  Table 42: Consultation Quality by Quartile of Experience and Cadre: Details (part II) Cadre Doctor C. Officer C. Assist Nurse Exp. Quartile 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Cough cough obs 8 4 1 14 12 15 10 21 23 24 30 21 1 6 16 15 duration of cough 1 1 0 .64 .58 1 .8 .9 .96 .96 .97 .62 0 .67 .81 1 sputum prod .5 0 0 0 .38 .57 .38 .19 .35 .53 .33 .31 0 .33 .25 .47 blood sput 0 0 0 0 .13 0 .33 .1 .06 0 .09 .18 0 0 .17 0 chest pain .63 0 0 .3 .25 .33 .13 .24 .12 .21 .38 .38 1 .5 .42 .4 diff breathing 0 .25 0 .07 0 .2 .1 .1 .13 .09 .73 .24 0 0 .19 .13 vaccinations 0 1 .17 0 0 0 .06 0 .06 .14 .07 0 0 .3 0 fever w/ cough .63 .75 1 .5 .64 0 .2 0 .59 .45 .07 .4 0 0 .25 0 cough: gen exam .75 0 0 0 .17 .73 .6 .57 .22 .58 .73 .57 0 .17 .88 .93 resp rate 0 0 0 0 0 .13 0 .24 .04 .13 .5 .05 0 .33 .19 .07 indrawing 0 0 0 .07 0 .07 .1 .29 .04 .3 .57 .1 0 .33 .31 .33 throat .38 0 0 0 .17 0 .2 .38 .04 .33 .4 .43 0 .5 .63 .53 auscultate .38 0 1 .86 .83 .53 .9 .67 .91 .83 .23 .76 0 .33 .88 .73 temp for cough .5 .5 1 .14 .42 0 .4 .14 .09 .05 .07 .1 0 0 .31 0 Diarr diarrhea obs 2 0 1 1 2 1 6 6 4 4 8 6 1 1 1 3 frequency .5 0 1 1 1 .83 .83 .75 1 .75 .5 1 1 0 .67 consistency 1 0 1 .5 1 .67 .67 .75 .5 .63 .67 1 1 1 .33 blood/mucus .5 0 1 .5 1 .33 .17 .5 0 .86 .17 1 1 0 0 vomiting 0 0 0 0 0 .83 .5 0 .5 .75 .5 1 1 1 0 fever w/ diarr 0 0 0 .5 0 .83 .5 .5 1 .75 .5 0 1 1 .33 diarr: gen exam .5 0 0 .5 1 .83 .5 0 1 .75 .5 1 0 0 .33 lethargy 0 0 1 0 1 .67 .5 .25 .75 .88 .67 1 1 0 .67 fontanella .5 0 0 0 0 .33 0 0 .5 .25 .5 0 1 0 0 pinch skin 0 0 0 0 .17 .4 0 .25 .5 0 0 1 0 0 weigh 0 0 0 0 1 .33 .67 0 .25 .88 .33 1 0 1 .33 temp for diarr 1 0 0 0 0 .33 0 0 .25 .25 .17 0 0 0 0 STD std obs 0 0 1 1 0 0 0 6 2 3 3 4 0 1 4 2 discharge/ulcer 0 0 .83 1 .67 .67 .75 1 .75 0 pain or itching 0 1 .2 0 .67 .67 .5 0 .75 0 fever w/ std 0 0 .33 0 0 0 0 0 0 .5 pain on urination 0 1 .5 0 .67 .67 .75 0 .25 .5 sexual history 0 0 .2 0 .33 0 .25 0 0 0 prev exposure 0 0 0 0 0 0 0 0 0 0 partners 0 0 0 0 0 .67 0 0 0 0 std: gen exam 0 0 .5 .5 .33 .33 .75 0 .5 1 skin rash 0 0 .33 .5 0 0 0 0 0 0 lymph nodes 0 0 0 0 .33 .33 0 1 .5 0 tenderness 1 0 .6 1 1 .67 .5 1 .75 0 genitalia 0 1 .67 1 .67 0 0 1 .5 0 precautions 0 0 .2 0 .33 .33 0 1 0 0 Other general obs 2 3 1 13 11 18 5 17 23 27 16 10 4 5 20 5 history to sympt 1 1 1 1 1 1 .83 .76 .93 .95 .88 .94 .83 .86 .89 .93 other: gen exam .5 0 0 0 .91 .67 0 .41 .39 .44 .31 .7 0 .4 .25 .8 exam to sympts .58 0 1 .5 .58 .6 .67 .71 .72 .64 .84 .39 .67 .71 .86 .67 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 43  Table 43: Consultation Quality by Quartile of Experience and Cadre: Total Scores Cadre Doctor C. Officer C. Assist Nurse Exp. Quartile 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 summary obs 14 6 2 20 17 27 27 46 36 45 70 51 3 8 40 42 o: open (avg) .41 -.77 .66 -.65 -.04 -.4 -.17 -.11 -.49 .22 .12 .1 .45 .72 .51 .52 o: expl (avg) -.36 -.89 -.3 -.81 -.86 -.18 -.79 .1 -.33 .07 .99 -.04 1.11 .68 .18 .42 o: close (avg) -.17 .5 .56 -.02 -.06 -.12 .38 -.13 .25 .19 .07 -.21 .17 .08 -.21 -.29 o: score (avg) -.11 -.55 -1.07 -.76 -.31 -.34 .11 .06 -.37 -.18 .46 .12 -.6 -.57 .25 .24 o: hist (avg) .07 -.39 -1.22 -.18 -.03 -.08 .28 -.19 .01 -.08 .43 .09 -.34 -.46 -.37 -.02 o: exam (avg) -.21 -.64 -.7 -.92 -.45 -.38 -.03 .25 -.57 -.11 .24 .1 -.48 -.48 .69 .37 Reported is the number of consultations observed and the fraction of each consultation that exhibited a particular procedure (if appropriate). 44  8 Consultation Quality (Using Vignettes) As should be apparent from the quality scores above, there are at least two problems with assessing quality by observing consultations. In the first place there are a series of facilities that have very few patients (or none at all). With so few observations we cannot authoritatively say we have seen anything representative of the facility. Second, the illnesses observed at one facility are very different from those observed at another, and therefore the things clinicians should do when diagnosing them are different. We have tried to correct for this problem by following four conditions and by normalizing scores before comparing them, but this is a crude correction. In close collaboration with the entire research team as well as Dr. Masatu and Dr. Jincen of CEDHA we developed a series of 6 vignettes or case study patients. In these vignettes we control for the symptoms presented because we have developed them ourselves. In addition, we can use the same evaluation at any facility whether or not there are patients. 8.1 The use of vignettes as a quality evaluation technique In order to assess certain levels of quality, we used vignettes or case study patients. In these cases we tried to imitate the consultation that might take place between a physician and a patient as closely as possible. We designed the case studies and knew what the correct diagnosis, treatment, and history taking should be. Therefore we were able to judge how well the clinician did compared to what was expected. All illnesses were specifically designed so that clinicians at any level of facility could properly diagnosis the condition. There was no condition for which a laboratory test was necessary to achieve correct diagnosis and no condition requiring medicines not available at all levels of facility. That does not mean that every level would respond in the same manner, but that a clinician at any level could achieve the perfect score. Each vignette and the instructions for administering the vignette can be found in Section A.2. Explanation of scores For each vignette the practitioner was judged against a list of questions that should be asked in history taking, a list of procedures that should be used to perform a physical examination and a list of points he should raise in health education. For each of these scores we derived two scores, a total score (v: hist (all) for history taking, v: exam (all) for physical examination and v: educ (all) for health education) and a score based on the most important of these questions (hist imp for history taking, exam imp for physical examination and educ imp for health education). The physician was compared to the average score for that vignette by normalizing the scores. Then all scores were summed. Thus each physician received a score based on his performance in 6 vignettes. In addition, we judged diagnosis, treatment use of lab tests and prescription by the following measures: