Skip to main content
. 2020 Oct 5;98(12):849–858A. doi: 10.2471/BLT.20.258145

Table 2. Independent variables, assessment of changes in health facility quality, United Republic of Tanzania, 2015–2020.

Construct and facility variable Definition of variable Data source Mean (SD)a
Outer setting
Patient needs and resources
  Population density No. of people within a 5 km radius of a health facilityb World population estimates for 201518 24 147 (63 733)
  Population demand for coveragec Percentage of women in district who gave birth in a facility in the past 5 years Demographic and Health Survey 201619 71 (22)
  Informed consumersc Percentage of women in district who completed primary education Demographic and Health Survey 2016 73 (15)
  Health-care agencyc Percentage of women in district who were involved in decisions about their own health care Demographic and Health Survey 2016 74 (13)
Cosmopolitanism
  Facility densityc Number of facilities in district per 100 000 population Star rating assessment 2015 15.4 (7.2)
  Urban councilc Percentage of facilities in town or municipal council areas and not in rural district council areas Star rating assessment 2015 12 (33)
Structural environment
  Accessibility Distance to major road, in km (bilinear interpolation)b OpenStreetMap 201620 2.32 (4.54)
  Remoteness Distance to city with a population of at least 50 000, in 10-km units Natural Earth II21 7.1 (5.1)
Peer pressure
  Facility rank at baseline Percentile rank of facility’s baseline star rating compared with other facilities in the same district Star rating assessment 2015 42 (33)
External policies and incentives
  Participated in results-based financing programme Percentage of facilities that participated in the results-based financing programme Star rating assessment 2015 15 (NA)
  Ineligible for results-based financing programme Percentage of facilities that were public facilities in a region participating in the results-based financing programme but had a baseline star rating of zero Star rating assessment 2015 11 (NA)
  Starter fund Percentage of facilities in an area eligible for a starter fund that had a baseline star rating of zero Star rating assessment 2015 4 (NA)
Inner setting
Structural characteristics
  Ownership Percentage of facilities that were public Star rating assessment 2015 81 (NA)
  Ownership Percentage of facilities that were private for-profit facilities Star rating assessment 2015 9 (NA)
  Ownership Percentage of facilities that were private non-profit facilities Star rating assessment 2015 10 (NA)
  Level Percentage of facilities that were dispensaries Star rating assessment 2015 85 (NA)
  Level Percentage of facilities that were health centres Star rating assessment 2015 12 (NA)
  Level Percentage of facilities that were primary-level hospitals Star rating assessment 2015 3 (NA)
  Baseline performance Facility star rating at baseline Star rating assessment 2015 0.81 (0.71)
Subsample analysis onlyd
External policies and incentives
  External supervision Percentage of facilities visited by an external supervisor in the previous 6 months who used a checklist, discussed facility performance and helped the facility make decisions based on data Service provision assessment22 2014–2015 76 (NA)
Structural characteristics
  Human resources Number of full-time health workers in each facilityb Service provision assessment 2014–2015 8.6 (21.9)
Culture
  Routine data use Percentage of facilities that reported routine use of a quality assurance system Service provision assessment 2014–2015 15 (NA)
  Client responsiveness Percentage of facilities with a procedure for reviewing patient feedback Service provision assessment 2014–2015 9 (NA)
  Community engagement Percentage of facilities that had a staff–community meeting within the previous 6 months Service provision assessment 2014–2015 64 (NA)
  Management function Percentage of facilities that acted after a recent management meeting Service provision assessment 2014–2015 46 (NA)

NA: not applicable; SD: standard deviation.

a All values are means and standard deviations unless otherwise noted.

b The natural log of this variable was used in the analytical models.

c Council-level variable.

d The subsample analysis included 672 facilities that took part in a service provision assessment between 2014 and 2015.