Background Non-communicable diseases and their risk factors are leading causes of disease burden in Iran and other
middle-income countries. Little evidence exists for whether the primary health-care system can eff ectively manage
non-communicable diseases and risk factors at the population level. Our aim was to examine the eff ectiveness of the
Iranian rural primary health-care system (the Behvarz system) in the management of diabetes and hypertension, and
to assess whether the eff ects depend on the number of health-care workers in the community.
Methods We used individual-level data from the 2005 Non-Communicable Disease Surveillance Survey (NCDSS) for
fasting plasma glucose (FPG) and systolic blood pressure (SBP), body-mass index, medication use, and
sociodemographic variables. Data for Behvarz-worker and physician densities were from the 2006 Population and
Housing Census and the 2005 Outpatient Care Centre Mapping Survey. We assessed the eff ectiveness of treatment on
FPG and SBP, and associations between FPG or SBP and Behvarz-worker density with two statistical approaches: a
mixed-eff ects regression analysis of the full NCDSS sample adjusting for individual-level and community-level
covariates and an analysis that estimated average treatment eff ect on data balanced with propensity score matching.
Results NCDSS had data for 65 619 individuals aged 25 years or older (11 686 of whom in rural areas); of these,
64 694 (11 521 in rural areas) had data for SBP and 50 202 (9337 in rural areas) had data for FPG. Nationally, 39·2%
(95% CI 37·7 to 40·7) of individuals with diabetes and 35·7% (34·9 to 36·5) of those with hypertension received
treatment, with higher treatment coverage in women than in men and in urban areas than in rural areas. Treatment
lowered mean FPG by an estimated 1·34 mmol/L (0·58 to 2·10) in rural areas and 0·21 mmol/L (–0·15 to 0·56) in
urban areas. Individuals in urban areas with hypertension who received treatment had 3·8 mm Hg (3·1 to 4·5) lower
SBP than they would have had if they had not received treatment; the treatment eff ect was 2·5 mm Hg (1·1 to 3·9)
lower FPG in rural areas. Each additional Behvarz worker per 1000 adults was associated with a 0·09 mmol/L
(0·01 to 0·18) lower district-level average FPG (p=0·02); for SBP this eff ect was 0·53 mm Hg (–0·44 to 1·50; p=0·28).
Our fi ndings were not sensitive to the choice of statistical method.
Interpretation Primary care systems with trained community health-care workers and well established guidelines can
be eff ective in non-communicable disease prevention and management. Iran’s primary care system should expand
the number and scope of its primary health-care worker programmes to also address blood pressure and to improve
performance in areas with few primary care personnel.