The importance of geographical distance to different types of R&D collaboration in the pharmaceutical industry

Figures & data

Table 1. Descriptive statistics of the pair between MNC R&D sites and external organizations with at least one collaboration.

	N	Mean	SD	Min	Max
Number of collaborations
Total	41 650	2.311	4.391	1	274
Basic science	5 721	1.864	2.411	1	72
Clinical science	37 759	2.273	3.945	1	202
Basic science-core	3 415	1.545	1.667	1	40
Basic science-exploration	3 497	1.571	1.711	1	40
Clinical science-core	26 155	1.943	3.101	1	162
Clinical science-exploration	21 280	1.667	2.157	1	106
Geographical distance (km)
Total	41 650	3 458	3 967	0	19 982
Basic science	5 721	2 986	2.411	0	18 805
Clinical science	37 759	3 461	3 955	0	19 982
Basic science-core	3 415	2 939	3 804	0	18 775
Basic science-exploration	3 497	3 216	3 939	0	18 805
Clinical science-core	26 155	3 358	3 861	0	19 454
Clinical science-exploration	21 280	3 665	4 081	0	19 982

Graph 1. Mean number of collaborations for different geographical distance intervals between MNC R&D sites and external organizations with confidence intervals. Geodist intervals (0–100 km, 100–500 km, 500–1 000 km, 1 000–2 000 km, 2 000–5 000 km, 5 000–8 000 km, 8 000–10 000 km, >10 000 km).

Table 2. Descriptive statistics of all possible pairs between R&D sites and external organizations.

	N	Mean	SD	Min	Max
Number of collaborations
Total	13 082 202	0.007	0.311	0	274
Total (excluding 0 km distances)	13 052 851	0.007	0.253	0	274
Basic science	13 082 202	0.001	0.121	0	72
Basic science (excluding 0 km distances)	13 077 907	0.001	0.070	0	72
Clinical science	13 082 202	0.010	0.256	0	202
Clinical science (excluding 0 km distances)	13 077 907	0.006	0.240	0	202
Basic science-core	13 082 202	0.001	0.035	0	40
Basic-core science (excluding 0 km distances)	12 879 785	0.001	0.032	0	40
Basic science-exploration	13 082 202	0.001	0.037	0	40
Basic-exploration science (excluding 0 km distances)	12 879 785	0.001	0.034	0	40
Clinical science-core	13 082 202	0.004	0.164	0	162
Clinical-core science (excluding 0 km distances)	12 879 785	0.004	0.160	0	162
Clinical science-exploration	13 082 202	0.003	0.112	0	106
Clinical-exploration science (excluding 0 km distances)	12 879 785	0.003	0.104	0	106
Geographical distance (km)
Total	13 082 202	5 965	4 294	0	19 982
Basic science	13 082 202	5 965	4 294	0	19 982
Clinical science	13 082 202	5 965	4 294	0	19 982
Basic science-core	13 082 202	5 937	4 294	0	19 982
Basic science-exploration	13 082 202	5 937	4 294	0	19 982
Clinical science-core	13 082 202	5 937	4 294	0	19 982
Clinical science-exploration	13 082 202	5 937	4 294	0	19 982

Table 3. Correlation matrix. *p < 0.05.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Y	1
2. LnGeodist	−0.02*	1
3. LnMass	0.12*	−0.025 *	1
4. Clin	0.01 *	0.000	0.000	1
5. Expl	−0.002*	0.000	0.000	0.00	1
6. Country	0.02 *	0.481*	0.013*	0.00	0.00	1
7. BP_Loc	0.006*	−0.047 *	0.040*	0.00	0.00	0.009*	1
8. Univ	0.013*	0.026 *	0.136 *	0.00	0.00	−0.025*	−0.044*	1
9. Hosp	−0.004*	−0.004 *	−0.043*	0.00	0.00	−0.074 *	−0.020*	−0.406 *	1
10. Big Pharma 1	0.011*	−0.018 *	0.104*	0.00	0.00	−0.018*	0.020*	0.127 *	−0.041*	1
11. Big Pharma 2	0.010*	−0.006 *	0.135 *	0.00	0.00	0.033*	0.019*	0.084 *	−0.040*	−0.047 *	1
12. Big Pharma 3	0.010*	0.046 *	0.123*	0.00	0.00	0.021*	0.007*	0.113*	−0.014*	−0.033 *	−0.058 *	1
13. Big Pharma 4	0.009*	−0.021 *	0.111*	0.00	0.00	−0.017*	0.007 *	0.083 *	−0.011*	−0.034*	−0.074 *	−0.058 *	1
14. Big Pharma 5	0.010*	0.008 *	0.141*	0.00	0.00	0.000*	0.013 *	0.067*	−0.020 *	−0.048*	−0.054 *	−0.089*	−0.083*	1
15. Big Pharma 6	0.011 *	0.032 *	0.097 *	0.00	0.00	0.004*	−0.003*	0.112*	−0.062 *	0.007 *	−0.028 *	0.030 *	−0.005 *	−0.025*	1

Table 4. Zero-inflated negative binomial regression model testing the number of collaborations of pharmaceutical MNCs for the period 2000–2012. Dependent variable: number of collaborations.

Negative binominal part	Model A (Estimate (SE))	Model B (Estimate (SE))	Model C (Estimate (SE))
(Intercept)	−1.610 (0.032) ***	−3.667 (0.053) ***	−4.161 (0.048) ***
LnMass $\left[{\alpha }_1\right]$	0.260 (0.002)***	0.253 (0.002) ***	0.242 (0.001) ***
LnGeodist $\left[{\beta }_1\right]$	−0.461 (0.003)***	−0.482 (0.007) ***	−0.483 (0.006) ***
Country$\left[{\beta }_2\right]$
Clin $\left[{\beta }_3\right]$		2.013 (0.057) ***	1.951 (0.047) ***
LnGeodist* Clin $\left[{\beta }_5\right]$		0.026 (0.008) **	0.035 (0.006) ***
Exp $\left[{\beta }_4\right]$			−0.545 (0.038) ***
LnGeodist *Exp $\left[{\beta }_6\right]$			0.050 (0.005) ***
BP_Loc $\left[{\gamma }_1\right]$	−0.118 (0.013)***	−0.123 (0.013) ***	−0.127 (0.011) ***
Univ$\left[{\gamma }_2\right]$	−0.033 (0.016)*	−0.035 (0.016) *	−0.055 (0.014) ***
Hosp $\left[{\gamma }_3\right]$	0.115 (0.013) ***	0.039 (0.013) ***	0.062 (0.012) ***
BigPharma1 $\left[{\gamma }_4\right]$	−0.209 (0.014) ***	−0.217 (0.013) ***	−0.196 (0.012) ***
BigPharma2 $\left[{\gamma }_5\right]$	−0.236 (0.012) ***	−0.218 (0.012)***	−0.194 (0.010) ***
BigPharma3 $\left[{\gamma }_6\right]$	−0.219 (0.013) ***	−0.216 (0.012) ***	−0.176 (0.010) ***
BigPharma4 $\left[{\gamma }_7\right]$	−0.252 (0.012) ***	−0.228 (0.012) ***	−0.196 (0.010) ***
BigPharma5 $\left[{\gamma }_8\right]$	−0.192 (0.012) ***	−0.188 (0.012) ***	−0.1545 (0.010) ***
BigPharma6 $\left[{\gamma }_9\right]$	−0.195 (0.021) ***	−0.161 (0.019) ***	−0.185 (0.017) ***
Zero-inflated part
(Intercept)	4.287 (0.035) ***	4.324 (0.035) ***	4.408 (0.032) ***
LnMass$\left[{\alpha }_1\right]$	−1.015 (0.015) ***	−1.036 (0.015) ***	−1.015 (0.011) ***
Fit statistics
Overdispersion (alpha)	11.764 ***	12.690***	10.124***
Log Likelihood	−24 070	−26 240	−32 180
Mc Fadden’s Adj. R2	0.295	0.322	0.342
Number of observation	13 082 202	26 164 404	52 328 808
Non zero observations	41 650	43 480	54 347

p *** <0.001, ** <0.01, *< 0.05.

Table 5. Zero-inflated negative binomial regression model testing the number of collaborations of pharmaceutical MNCs for the period 2000–2012. Dependent variable: number of collaborations.

Negative binominal part	Model D (Estimate (SE))	Model E (Estimate (SE))	Model F (Estimate (SE))
(Intercept)	−3.416 (0.038) ***	−5.357 (0.057)***	−5.752 (0.052) ***
LnMass $\left[{\alpha }_1\right]$	0.262 (0.002) ***	0.253 (0.002) ***	0.242 (0.001) ***
LnGeodist $\left[{\beta }_1\right]$	−0.283 (0.004) ***	−0.320 (0.007) ***	−0.336 (0.007)***
Country$\left[{\beta }_2\right]$	1.161 (0.015) ***	1.142 (0.015) ***	1.089 (0.012) ***
Clin $\left[{\beta }_3\right]$		1.912 (0.055) ***	1.873 (0.047) ***
LnGeodist* Clin $\left[{\beta }_5\right]$		0.041 (0.007) ***	0.048 (0.006) ***
Exp $\left[{\beta }_4\right]$			−0.557 (0.037) ***
LnGeodist *Exp $\left[{\beta }_6\right]$			0.044 (0.005) ***
BP_Loc $\left[{\gamma }_1\right]$	−0.039 (0.013) ***	−0.045 (0.013) ***	−0.053 (0.011) ***
Univ$\left[{\gamma }_2\right]$	0.055 (0.016) ***	0.055 (0.016) ***	0.037 (0.014) **
Hosp $\left[{\gamma }_3\right]$	0.221 (0.013) ***	0.147 (0.013) ***	0.175 (0.011) ***
BigPharma1 $\left[{\gamma }_4\right]$	−0.169 (0.014) ***	−0.178 (0.013)***	−0152 (0.011) ***
BigPharma2 $\left[{\gamma }_5\right]$	−0.269 (0.012) ***	−0.248 (0.012) ***	−0.220 (0.010) ***
BigPharma3 $\left[{\gamma }_6\right]$	-−0.249 (0.013) ***	−0.246 (0.012) ***	−0.210 (0.010) ***
BigPharma4 $\left[{\gamma }_7\right]$	−0.207 (0.012) ***	−0.185 (0.012) ***	−0.152 (0.010) ***
BigPharma5 $\left[{\gamma }_8\right]$	−0.176 (0.012) ***	−0.173 (0.012) ***	−0.141 (0.010) ***
BigPharma6 $\left[{\gamma }_9\right]$	−0.272 (0.020) ***	−0.236 (0.020) ***	−0.244 (0.017) ***
Zero-inflated part
(Intercept)	4.313 (0.035) ***	4.324 (0.035) ***	4.424 (0.033) ***
LnMass$\left[{\alpha }_1\right]$	−1.062 (0.015) ***	−1.036 (0.015) ***	−1.048 (0.013) ***
Fit statistics
Overdispersion (alpha)	10.836***	13.112***	9.431***
Log Likelihood	−23 777	−25 940	−31 800
Mc Fadden’s Adj. R2	0.303	0.322	0. 342
Number of observation	13 082 202	26 164 404	52 328 808
Non zero observations	41 650	43 480	54 347

p *** <0.001, ** <0.01, *< 0.05.

Table A1. Classification of the articles into therapeutic areas through the heading MeSH term.

Therapeutic Area	Heading MeSH Term
Infectious diseases	Virus Diseases [C02]
Musculoskeletal diseases	Musculoskeletal Diseases [C05]
Endocrinology	Endocrine System Diseases [C19]
Cardiovascular diseases	Cardiovascular Diseases [C14]
Neuroscience	Nervous System Diseases [C10]
Ophthalmology	Eye Diseases [C11]
Oncology	Neoplasms [C04]
Respiratory diseases	Respiratory Tract Diseases [C08] Otorhinolaryngology Diseases [C09]
Immunology	Immune System Diseases [C20]

Figure A1. Density of the distribution of collaborations according to the distance (km) between an external organization and an MNC R&D site. Kernel density graphs.